The choice of the ‘best’ or ‘preferred’ keyboard instrument in eighteenth-century European music has been a favourite subject of discussion for scholars as well as performers. The problem is inextricably connected to the rapid evolution during that time of keyboard instruments and technique, or rather of what might better be described as keyboard ‘idiom’. These matters have been considered with particular enthusiasm by students of the music of Carl Philipp Emanuel Bach and Joseph Haydn, often in connection to the question of how Bach might have influenced his younger contemporary.Footnote 1
No consensus has ever been reached on these issues, perhaps because the questions asked have not been exactly the right ones. The present investigation reframes the matter by first considering the concept of ‘keyboard idiom’, then examining the keyboard idiom of individual compositions and ending with a reconsideration of the musical relationship between the two composers. The chief finding is that, although both musicians probably did their actual music-writing at the clavichord, neither intended their keyboard music for specific types of instruments until much later in their careers than is usually thought. Rather, they conceived most of their keyboard works for a generic ‘clavier’, even if individual movements seem to favour one type of instrument (such as the clavichord) or another. Even relatively late keyboard works, both solo and accompanied, are playable – and were played – on various types of harpsichord. Each composer began to write for a more specific type of keyboard instrument, that is, some variety of the fortepiano, only when dynamic effects and a new type of musical rhetoric became essential elements of compositional thought.
Instrument and Idiom
Eighteenth-century Europe knew many types of keyboard instruments: not only organs, harpsichords, clavichords and pianos, but also rarer and more exotic varieties.Footnote 2 None of these had a standard form; when today we use an expression like ‘clavichord’ or ‘fortepiano’, we refer only to an abstraction defined by a certain general type of keyboard mechanism. The possibility of controlling dynamics through the player's touch was an obvious distinguishing feature of the newer types of keyboard instrument. But just as critical as touch sensitivity is the overall range of dynamics. Equally important is sustaining power, from which derives the possibility of playing a legato melody in long notes. Sonority is also significant, whether determined by distinct stops or registrations, modified by the use of dampers and other devices, or varied simply through contrasts in sound between different tessituras on a single keyboard.
Dynamics are the easiest of these features to discuss, if only because they can be discretely notated in a score. There are, however, at least three distinct types of dynamic marking in eighteenth-century keyboard music. The so-called terraced dynamics of alternating forte and piano already occur in pieces for a two-manual harpsichord, such as J. S. Bach's Italian Concerto. Gradual change in dynamic level, that is, crescendo and diminuendo, could be indicated by the words themselves, by hairpin symbols or by the placement of individual signs for piano and forte in a way that implies a gradual change from one to the other. Representing a third type of dynamic marking are those isolated signs for forte or sforzando that start to appear around 1760 to indicate a momentary accent within a relatively quiet passage.
How a player responds to each type of dynamic sign, if at all, depends on the instrument, and it does not necessarily involve touch alone. Nor does the presence of dynamic markings in a keyboard score necessarily indicate a specific intended instrumental medium. Three levels of dynamics, such as forte, piano and pianissimo, already occur in Bach's Second ‘Prussian’ Sonata, Wq48/2, of 1740.Footnote 3 What this implies about choice of instrument, in a publication that was issued ‘per Cembalo’, is not entirely clear. It hardly indicates abandonment of the harpsichord, for on a two-manual instrument one can divide the hands between louder and softer keyboards to express an intermediate dynamic level.Footnote 4 Eighteenth-century harpsichords could incorporate Venetian swells, pliable leather plectra and machine stops, each permitting a certain degree and type of dynamic inflection.
More fundamentally, composers did not necessarily expect every dynamic sign to be realized literally. Well into the nineteenth century, composers continued to write unrealizable dynamic indications, as in several famous instances of hairpins marked on single chords.Footnote 5 During Haydn's later years, composers were calling for the damper pedal in passages that include rests and staccatos, which therefore no longer signify an actual cessation of sound; rather, the hand springs away from the keyboard as the notes continue to ring.Footnote 6 In short, there was a tradition that the notation of music for keyboard instruments did not represent what one actually heard. Dynamic indications, in particular, must have been regarded much like ornament signs, the realization of which might be desirable but not essential to the composition. This was implicit in the publication of keyboard music that was described as being for either piano or harpsichord, a player of the latter not being expected to realize every dynamic indication. One might imagine that composers preferred to hear every dynamic sign realized in sound. Yet it cannot be necessarily assumed that as Bach, Haydn and their contemporaries wrote increasingly numerous and detailed dynamic indications into their music, they took use of the piano for granted. In any case, dynamics are only one element of keyboard idiom.
Keyboard idiom in works by C. P. E. Bach
‘Keyboard idiom’ means how the composer writes for the keyboard, that is, the uniquely clavieristic gestures and textures of the music. To speak of musical textures and gestures is to speak metaphorically; behind those metaphors stand rather fuzzy concepts that are difficult to define rigorously. Still, we might posit that what makes a particular composition ‘idiomatic’ to a given type of instrument is writing that exploits its unique features. Identifying such features is, to some degree, subjective, but four examples may illustrate the evolving keyboard idiom in works by Emanuel Bach. One of these is idiomatically generic; the others are increasingly specific to particular types of keyboard instrument. All four are from published works that Haydn could have known (Example 1a–d).
C. P. E. Bach: (a) Fugue in E flat major, Wq119/6, bars 119–123, from Clavierstücke mit einem practischen Unterricht für Anfänger und Geübtere, Zweyte Sammlung (Berlin: Haude und Spener, 1762); (b) ‘Württemberg’ Sonata No. 6 in B minor, Wq49/6/i, bars 1–5, from Sei sonate per cembalo (Berlin: Giovanni Ulrico Haffner, 1744); (c) Sonata in C minor Wq50/6, bars 9–20, from VI. Sonates pour le clavecin avec des reprises variées (Berlin: George Louis Winter, 1760); (d) Rondo in A minor, Wq56/5, bars 21–28, 56–57, from Clavier-Sonaten nebst einigen Rondos fürs Forte-Piano für Kenner und Liebhaber, Wq56 (Leipzig: author, 1780)
Example 1a is from a fugue composed in 1755 and published several years later with analytical commentary by Friedrich Wilhelm Marpurg. A recent edition places this composition in a volume of organ music, but it lacks a pedal part, and Marpurg's readers are likely to have studied it at home at a stringed keyboard instrument. It is in the same quasi-vocal or pseudo-renaissance style as J. S. Bach's Art of Fugue, which had appeared in print with Marpurg's preface a decade earlier. Example 1b is from the ‘Württemberg’ Sonatas, published in 1744.Footnote 7 It juxtaposes grand arpeggiated chords with quieter music; this is typical of writing for a two-manual harpsichord. Example 1c, from the Reprisen-Sonaten of 1759, shows a proliferation of dynamic markings but also the thin, treble-dominated texture that is especially favourable for the clavichord.Footnote 8
Example 1d is from a rondo explicitly for the fortepiano. The designation of instrument is included in the title of the original publication: Clavier-Sonaten nebst einigen Rondos fürs Forte-Piano für Kenner und Liebhaber (Leipzig: author, 1780). Although the syntax might be ambiguous, the volume was the second in a series of six, and in later volumes the layout and typography of the printed title-page leave little doubt that the words ‘fürs Forte-Piano’ refer specifically to the volume's three rondos (Figure 1). Printed title-pages might reflect only the publisher's notion of what might make for a marketable product. In this case, however, the close relationship between composer and publisher (as documented by correspondence) makes it unlikely that Bach would have objected to the wording or layout of the title-page, which underwent no substantive changes in the remaining three issues in the series.
Bach, Clavier-Sonaten und Freye Fantasien nebst einigen Rondos fürs Fortepiano für Kenner und Liebhaber, Wq58 (Leipzig: author, 1783), title-page
Composed in 1778, the rondo shown in Example 1d substantially expands the composer's vocabulary of keyboard gestures. These now include crescendi and diminuendi, which may (as in bars 25–26) accompany chromatic scale fragments in octaves. In addition, there is arpeggiated passagework (bars 56–63), which looks generic but sounds especially striking if played on a piano without dampers.Footnote 9 There are good reasons for doubting whether individual gestures can be specifically indicative of one keyboard instrument or another.Footnote 10 Yet the crescendo marking beneath chromatic octaves for the right hand shown in Example 1d, within a composition characterized by numerous close-spaced dynamic indications, supports the implication of the title-page that this example was written with some sort of piano in mind – unlike the three earlier passages illustrated.
Keyboard idiom in music by Haydn
Haydn's keyboard works, likewise written over a span of half a century, reveal a comparable range of idioms. Like those of C. P. E. Bach, Haydn's compositions for keyboard have long been studied for clues regarding either their intended or their most effective sounding medium. Reading the recent literature on the subject, one can gain the impression that each piece must have been written for some particular instrument, if only we could determine which.Footnote 11 The two most searching investigators of the topic for Haydn, A. Peter Brown and Bernard Harrison, acknowledged the problematic nature of the enterprise. They nevertheless sifted through various types of evidence – the availability of instruments in Vienna, letters and portraits that mention or depict keyboards, dynamic markings and titles in published music – to trace the composer's transition to writing for ‘fortepiano’. Indeed, Brown provided a table that provisionally indicates the ‘preferred instrument’ and ‘other possible instrument’ for every solo and accompanied sonata.Footnote 12
Yet neither a printed title-page specifying an instrument, nor a remark in a letter favouring a keyboard of a particular type or by a particular maker, can prove that a composer thought in terms of matching an individual piece with a specific instrument – or even with a general type. Haydn, like his younger contemporaries, seems to have taken the piano for granted by the end of his career. How and exactly when he began to do that remain obscure. Moreover, ‘piano’ or ‘fortepiano’ remains a vague formulation, given the tremendous variety of instrument types available throughout Haydn's lifetime, even within a single major city.
From generic ‘clavier’ to piano
Harrison argued that Haydn's compositional evolution was less a matter of shifting from one specific keyboard type to another, than a transition from a ‘generalized keyboard idiom’ to ‘writing for a specific keyboard instrument’.Footnote 13 This is surely right, but does ‘a specific keyboard instrument’ refer to pianos generally (as opposed to harpsichords or clavichords), a particular type of piano (say, a Viennese square) or one particular instrument owned by the composer or a dedicatee (such as Therese Jansen, for whom Haydn wrote four of his late piano trios)? Haydn, like his contemporaries, could only gradually have come to understand that particular keyboard instruments possess particular qualities that can be exploited compositionally. Today this may seem self-evident, yet most of Haydn's early keyboard works, like those of Bach, are not distinctly idiomatic even to a general category of instrument. That same sixth ‘Württemberg’ Sonata whose opening movement seems so clearly destined for a two-manual harpsichord concludes with an extended two-part invention. This movement (see an excerpt in Example 2), which lacks dynamic indications, is as generic, with respect to keyboard idiom, as a fugue in stile antico. Both Bach and Haydn continued to write movements until at least the 1760s that are confined to this same plain two-part texture. The figuration lies well under the hands, but it calls for few if any dynamics, and it does not obviously reflect any thought as to whether it would be best played on organ, harpsichord or clavichord – or, for that matter, by flute or violin together with bassoon or cello.
Bach, ‘Württemberg’ Sonata No. 6 in B minor, Wq49/6/iii, bars 1–7
In Vienna, Matthias Georg Monn, Georg Christoph Wagenseil and Joseph Anton Steffan employed a similarly generic idiom, at least in their earlier keyboard works, continuing the practice of Italian predecessors such as Domenico Alberti and Domenico Scarlatti. Scarlatti, at least, had access to pianos, yet there is nothing in his keyboard music that demands a touch-sensitive instrument.Footnote 14 The key word is demand. One can use a clavichord, fortepiano or for that matter a modern piano to add dynamics and colour to earlier keyboard music. But dynamics and colours are embellishments, and not necessarily welcome ones, in music composed without consideration of them. László Somfai argued more than twenty years ago for the use of Viennese-style harpsichords in Haydn's keyboard music composed before 1780.Footnote 15 Such instruments typically have a single manual, often with a distinctive variety of short or ‘broken’ bass octave.Footnote 16 Yet one continues to hear these compositions played on fortepianos – often on varieties that became prevalent only after 1790, like the Walther instrument owned by Wolfgang Amadeus Mozart but modified by (probably) another maker at an uncertain date.Footnote 17 Such an instrument may be perfectly suited to the edited versions of Haydn's music that began to come out during the same period. But even if these editions reflect the composer's own way of playing on pianos of the late eighteenth century, they do not necessarily correspond with earlier practice. For this reason they may be misleading with respect to how the composer came to write as he did for the keyboard.
Few harpsichordists today regularly perform solo keyboard compositions by Haydn, J. C. Bach or Mozart. Doing so requires some departure from the type of playing that has become customary for older music. Yet the French style of two-manual harpsichord that is now so often used for baroque repertory is in fact a classical type, most modern examples being copies of a handful of instruments made or modified in Paris in the 1760s and 1770s.Footnote 18 These instruments were designed for playing music by the contemporaries of Haydn, J. C. Bach and Mozart, even if all three composers were tending towards the piano during the period. Doubtless both harpsichords and pianos were used for playing their music, but whether either type of instrument is required for specific compositions is an entirely different matter.
Richard Maunder found details in Haydn's six sonatas published in 1774, with a dedication to his patron Nikolaus Esterházy (hXVI:21–26), that, taken literally, point to the need for a two-manual harpsichord. Yet he also provided evidence that such instruments were rare in Vienna at the time. Even in a sonata published two years later, there is, in fact, nothing particularly suggestive of a two-manual instrument in a passage that Maunder described as ‘perversely awkward on a single-manual instrument’ (Example 3).Footnote 19 Rather, the interlocking of the two hands required here is one of those clever but not really virtuoso keyboard techniques that Haydn seems to have cultivated. It is one of the elements in his distinctive approach to the keyboard that also included large leaps (sometimes requiring hand-crossing) but nothing unusually difficult, awkward (‘perversely’ or otherwise) or obviously calculated for a particular instrument. Only from 1790 do we have a document in which Haydn says he is no longer accustomed to ‘writing for the harpsichord’.Footnote 20 Like other composers of the time, he apparently had always been in the habit of composing at the clavichord, but he did not necessarily regard that instrument as a desirable or optimal medium for any particular composition.
Joseph Haydn, Sonata in E flat major hXVI:28/i, bars 105–110, from Six sonates pour le clavecin ou le piano forte, Op. 14 (Berlin and Amsterdam: Hummel, 1778)
Hence, although we know something about the types of instruments owned by both Bach and Haydn, such information provides little guidance for how we might interpret individual compositions.Footnote 21 Other composers who worked at the clavichord include Johann Baptist Wanhal, Giuseppe Bonno (Kapellmeister at the Viennese court), Mozart and Steffan. Charles Burney reported that Wanhal played for him ‘six lessons which he had just made for that instrument’.Footnote 22 That was in 1772, before any of Wanhal's keyboard music had appeared in print – none of it designated as clavichord music. Nor are there any titles or other sources referring to use of the instrument in Steffan's keyboard music; the assertion that his early published sonatas ‘were composed for the fortepiano’ apparently depends on doubtful assumptions about the ‘limited volume’ of the clavichord.Footnote 23 An argument that some of the same music ‘seems to work particularly well on the clavichord’ is little more than an expression of personal preference.Footnote 24
Bach, eighteen years older than Haydn, described the clavichord as the instrument on which ‘a keyboard player can be judged most conveniently’.Footnote 25 That he liked to play the clavichord is clear from accounts by visitors, including Burney and the poet Matthäus Claudius. Yet he rarely designated individual pieces for specific keyboard instruments, even though there are reasons for supposing that, by the 1750s, his solo keyboard music was written primarily for the clavichord. Nevertheless, the late rondos, as we have seen, were designated for the fortepiano, which he also used in public performances of concertos and other compositions after his move to Hamburg in 1768.Footnote 26
Something similar appears to have been true as well of Emanuel's younger half-brother Johann Christian. This Bach studied with his older sibling in Berlin for five years before departing in 1755 for Italy and ultimately England. During his time in Italy, when he was employed as an organist at Milan, he must otherwise have played almost exclusively the harpsichord. Yet his English keyboard music includes compositions evidently written for piano, as implied not only by the presence of dynamic indications but by elements of texture discussed below in connection with Haydn. Richard Maunder argued that Johann Christian Bach had adopted the piano for public performances by the early 1770s.Footnote 27 Dynamic markings, including closely spaced ‘p’ and ‘f’ and a single ‘cresc’, appear as early as in the first three sonatas of his Op. 5 (London: Welcker, c1766), the title-page of which declared it to be ‘for the Piano Forte or Harpsichord’. These works could have been known in Vienna by early 1772, when they were advertised as sonatas for harpsichord or piano.Footnote 28
Haydn, nevertheless, stands apart from these north European composers and probably also from Mozart (who was more widely travelled at a younger age and more of a keyboard virtuoso). Mozart, inspired by Johann Christian Bach, was probably writing for the piano by the time of his six accompanied sonatas published in Paris in 1779.Footnote 29 The following year saw the publication of the rondo by Emanuel Bach shown in Example 1d, explicitly for piano, and around the same time there appeared Haydn's Op. 30 sonatas, hXVI:35–39 and 20. Several of these are sometimes seen as among Haydn's first sonatas to call distinctly for the piano.Footnote 30 Indeed, within the first sonata one finds four dynamic levels, from pianissimo to fortissimo. Also present in this sonata (hXVI:35) is the same chromatic crescendo in octaves present in Emanuel's rondo; Haydn marks this with a slur (Example 4). A variation of this idea occurs in the C sharp minor sonata (hXVI:36) from the same collection (Example 5). Mozart, incidentally, employed the same gesture – albeit shared between the two hands – in his first published sonata for solo keyboard (k309 (284b)), which appeared about a year later (Example 6).Footnote 31
Haydn, Sonata in C major hXVI:35/i, bars 130–136, from Sei sonate per il clavicembalo, o forte piano, Op. 30 (Vienna: Artaria, 1780)
Haydn, Sonata in C sharp minor, hXVI:36/i, bars 20–23, from Sei sonate per il clavicembalo, o forte piano
Wolfgang Amadeus Mozart, Sonata in C major k309 (284b)/i, bars 107–111, from Trois sonates pour le clavecin ou le forte piano, Op. 4 (Paris: Heina, c1781)
Even without the dynamic markings, this gesture is peculiarly pianistic. To produce the illusion of legato, the pianist relies not only on the sustaining power of the instrument but also on the possibility of precisely matching the dynamic level of each successive tone with that of the preceding one – or of producing an unbroken crescendo or diminuendo. It also helps not to hear the articulation of every note produced by the plucking mechanism of the harpsichord or the striking of the tangents on the strings of a clavichord. To be sure, the hammers of some early pianos can also create the impression of articulating every note, and the illusion of legato can be approximated on a lightly quilled harpsichord or a clavichord. But even the most resonant clavichord can produce only a feeble crescendo on this figure, and the same is quite beyond the capabilities of any ordinary harpsichord.
Another feature of many of these same pieces – even Haydn's Sonata in C minor – is the frequent use of the so-called Alberti bass (Example 7a). Today this is associated with early piano music. Yet Alberti himself, who died in 1746, used his signature device in many sonatas that he certainly expected to be played on Italian harpsichords with a single manual (Example 7b). Modern pianists usually play Alberti basses legato, even applying the damper pedal, and indeed legato performance is documented in eighteenth-century sources.Footnote 32 Yet the Alberti bass often imitates textures from orchestral and ensemble music, in which the broken chords are likely to have been played lightly and detached, as in a string trio by J. C. Bach (Example 8). Execution in a similar manner at the keyboard makes it easier for the player to project a sustained melody against the accompaniment. This is true even for slow movements, as in hXVI:33 (Example 9a), although the effect would be more visceral in the scampering finale of hXVI:34 (Example 9b); both illustrations are from Haydn's sonatas of c1780.Footnote 33 As pianistic as this notation may appear today, when newly composed, the music using it must have seemed perfectly suited even to a one-manual harpsichord, as indeed it is when the player takes advantage of that instrument's capability for clear articulation.
(a) Haydn, Sonata in C minor, hXVI:20/i, bars 33–35, from Sei sonate per il clavicembalo, o forte piano; (b) Domenico Alberti, Sonata per cembalo in C major/i, bars 1–2 (Bologna, Museo Internazionale e Biblioteca della Musica di Bologna, CC.159/3)
Johann Christian Bach, Trio in D major for two violins and cello, WarbB36, first movement, bars 1–6, from Six sonates pour deux violons et basso (Amsterdam, 1772)
Haydn, (a) Sonata in D major hXVI:33/ii, bars 14–16, from Trois sonates pour le clavecin ou piano-forte, Op. 46 (Paris: LeDuc, c1785); (b) Sonata in E minor hXVI:34/iii, bars 1–4, from Œuvres complettes, volume 1 (Leipzig: Breitkopf und Härtel, 1799)
Pieces for Specific Instruments?
In addition to the late rondos for piano, a handful of Bach's compositions can be assigned to particular types of instruments. One sonata composed in 1747 (Wq69) was for a special harpsichord for which the composer specified registrations. Several other pieces contain indications for Bebung, the vibrato-like ornament distinctive to the clavichord. Six or seven sonatas are designated as organ works, although they lack pedal parts. All of these compositions are equally playable on other instruments. Nevertheless, one can detect a trend in Bach's music for solo keyboard from the 1740s onward towards a type of writing that is distinctly favourable for the clavichord, even when explicit dynamic markings are absent. These works are typified by thin textures, or a melodic line accompanied by simple chords, with numerous appoggiaturas and other expressive ornaments in both treble melody and bass.Footnote 34
A few of Haydn's earlier keyboard pieces, especially ones not destined for immediate publication, might likewise have been composed with a specific type of instrument in mind. For instance, the Capriccio in G major (hXVII:1) is playable as written only on a keyboard with a Viennese broken octave, and it entirely lacks dynamic markings. It is not easy for a performer on the harpsichord to bring out the main theme when it appears in the middle register, accompanied by an Alberti ‘bass’ in the treble (Example 10). Yet Haydn is unlikely to have had the piano in mind at the early date indicated by the autograph (1765).Footnote 35 As the piece also avoids idioms particularly characteristic of the clavichord (such as in Example 1c, or in Example 11 below), a Viennese one-manual harpsichord is the most plausible ‘intended’ instrument – albeit only in the sense that this was the medium in which the composer might have expected it to be played most often. Even if the passage illustrated is not particularly idiomatic to the harpsichord, it hardly demands a touch-sensitive instrument.
Haydn, Capriccio in G major, hXVII:1, bars 233–237, from Œuvres complettes, volume 2 (Leipzig: Breitkopf und Härtel, 1800)
Instances of Bebung from Bach: (a) Probestück No. 6, Wq63/6/iii (fingerings omitted), from the Exempel nebst achtzehn Probe-Stücken in Sechs Sonaten issued alongside volume 1 of the Versuch über die wahre Art das Clavier zu spielen (Berlin: Christian Friedrich Henning, 1753); (b) Sonata in F major Wq55/2/ii, bars 69–74, from Sechs Clavier-Sonaten für Kenner und Liebhaber, Wq55 (Leipzig: author, 1779)
Fifteen years later Haydn was probably still thinking along the same lines. Even the six sonatas Op. 30, dedicated to the Auenbrugger sisters and published by Artaria about 1780, seem less clearly destined for the piano than is sometimes supposed. The keyboard idiom has evolved, growing more intricate and variable than that found in the Capriccio, yet it remains problematical to see in these pieces – even the most famous one, discussed below – demands for a specific type of instrument. The previous year, 1779, had seen the publication in Vienna of a treatise on clavichord playing by Franz Rigler, which was supplemented by six ‘clavichord pieces of various types’.Footnote 36 The latter phrase echoed the title of an earlier publication by Bach, the Clavierstücke verschiedener Art, Wq112 (Berlin: Georg Ludwig Winter, 1765). Rigler makes it clear that for him, as for Bach, the word ‘Clavier’ referred primarily to the clavichord, which must have been found in many homes in Vienna. Nevertheless, Rigler says nothing specifically about dynamics, suggesting that, even when playing the clavichord, musicians considered dynamics secondary in importance to ornaments and other elements of performance.Footnote 37
Haydn's Sonata in C minor – for clavichord?
Among Haydn's earlier keyboard pieces, the one most often regarded as calling for a specific instrument – the clavichord – is the Sonata in C minor, hXVI:20. That the composer recognized its special character might be inferred from the fact that he published it only as the final item in the 1780 collection, a decade or so after its composition.Footnote 38 The view of this sonata as specifically, and perhaps uniquely, for clavichord reflects not only its minor key but other features. Within the 1780 collection, it is exceptional for its numerous dynamic markings as well as its varied textures, which tend, especially in the slow movement, toward sustained linear writing for both hands.Footnote 39
Yet, as in the other sonatas with which it was published, many passages employ the Alberti bass (see Example 7a). The sonata lacks any notation for Bebung, for which, however, even Bach called explicitly in only a handful of pieces, including two notable ones that Haydn is likely to have known (Examples 11a and 11b).Footnote 40 Haydn never uses this device, although the similar notation for the Tragen der Töne (a sort of portato) does occur in some later compositions, such as the piano trio hXV:22 (Example 12a and 12b). This trio, incidentally, also uses Bach's tenuto indications, as well as the irrational division of the beat which Bach described as tempo rubato.Footnote 41 For Haydn, however, the Tragen der Töne cannot have been a unique signifier for the clavichord.
Instances of Tragen der Töne from (a) Bach, Probestück No. 6, Wq63/6/iii; (b) Haydn, Piano Trio in E flat major hXV:22/ii, bars 17–19 (strings omitted), from Trois sonates pour le piano forte avec accompagnement de violon & violoncello, Op. 71 (London: Preston and Son, 1795)
The tonality of the C minor sonata, especially the slow movement in A flat major, makes it relatively difficult to play well on any clavichord. This is because the short accidental keys of most such instruments make it challenging to sustain tones and play legato in tonalities with many accidentals. Moreover, this sonata contains several passages whose performance is problematic on a so-called fretted clavichord. Such an instrument – the earliest, cheapest and historically most widespread type – has multiple pitches assigned to each string. For example, F and F♯ (or G♭) might be produced by tangents striking the same string at different points.Footnote 42 ‘Fretting’ reduced the size and cost of the instrument but also made it impossible to play a descending slur between the notes in question, which also could not be struck together in a chord or alternated rapidly in a trill.
Use of a fretted instrument would affect the performance of passages such as the slurred chromatic melisma for the right hand in the middle of the first-movement exposition (Example 13). A keyboard player normally produces legato by means of an imperceptible overlap between two notes. This becomes impossible if the first note must be detached to allow the second to be struck on the same string. Failure to detach the first note on a fretted clavichord leads to a ‘blocked’ tone (or just a thump) instead of two notes joined by a slur. To be performed satisfactorily on a clavichord, this sonata would require a large unfretted instrument (that is, with one string per key) of the type that might have been common at this date only in northern Germany.Footnote 43
Haydn, Sonata in C minor, hXVI:20/i, bars 19–22
Maunder found substantial evidence for unfretted instruments in Vienna, but not before the 1780s.Footnote 44 To be sure, the relatively low value of even a large clavichord, as compared to a harpsichord or piano, would have made such instruments less likely to be advertised for sale or to leave traces in inventories and other documents. Rigler, whose treatise evidently was conceived as an elementary version of Bach's Versuch, must have assumed the use of a large instrument – to judge from the five-octave range required by the keyboard pieces in the musical supplement. Supporting the same conclusion is Rigler's provision of twenty-four model cadenzas in all keys. On the other hand, only one of Rigler's Clavierstücke employs a key signature with more than three flats or sharps. The latter occurs in a single minore within a rondo movement, and this and other pieces in the volume avoid the descending chromatic half steps so crucial in Haydn's sonata.
The clavichord or clavichords that Haydn reportedly owned in his early years may well have been small, fretted instruments, unsuited for performing much of the music presumably composed on them.Footnote 45 The Sonata in C minor contains no actual simultaneities that would have been unplayable on such instruments. Many slurs and ornaments, however, involve adjacent chromatic notes, as in Example 13. No composer who knew this type of clavichord, as Haydn undoubtedly did, could have expected this sonata to be performed on it exactly as written. He might have delayed publishing it for precisely that reason.
Haydn nevertheless could have found inspiration for this composition, including the key of its slow movement, in the last of the six sonatas that Bach published as a musical supplement to his Versuch. Footnote 46 Indications for Bebung in the concluding movement of that sonata, the ‘Hamlet’ fantasia, suggest that it was intended primarily for the clavichord (see Examples 11a and 12a above). Yet that same sonata includes a slow movement in A flat major, as well as an opening Allegro in F minor in which one hand repeatedly crosses over the other (Example 14). This is the only composition by Bach to exploit that technique, which we tend to identify with the sonatas of Domenico Scarlatti. Even so, the fact that Bach incorporated it into a composition primarily for the clavichord confirms Burney's later report that Bach was prepared to play ‘in every style’ on that instrument.Footnote 47 The last of the Probestück sonatas provided practice for any player to do the same, and Haydn might have intended his Sonata in C minor to serve a similar purpose.
Bach, Probestück No. 6, Wq63/6/i, bars 1–6 (fingerings omitted)
To be sure, Bach played chiefly in what Burney called the ‘expressive’ style. We might identify the latter as involving slow to moderate tempos in a thin treble-dominated texture, incorporating the dynamic gradations that Quantz as well as Bach described as contrasting ‘shadow and light’ (‘Schatten und Licht’).Footnote 48 The first two movements of Haydn's C minor sonata surely fall within this category, which corresponds to ideas of what is today considered appropriate to the clavichord. But the same could not be said of the last movement, which reaches a climax in a long hand-crossing passage, an extract from which is shown in Example 15. Even Beethoven might have taken an idea or two from Bach's piece in his Pathétique sonata of 1798. Yet that is surely not clavichord music, even if it remains playable on a five-octave instrument – and at least one early edition still described it as being for ‘harpsichord or piano’ (Example 16).
Haydn, Sonata in C minor, hXVI:20/iii, bars 108–111
Ludwig van Beethoven, Sonata Pathétique, Op. 13/i, bars 220–225. Grande sonate pathétique pour le clavecin ou piano-forte (Vienna: Eder, c1800)
These examples show how an idiom that originated in harpsichord music could be transferred to the clavichord and the piano. Although Haydn's C minor sonata benefits from resources beyond those of a typical Viennese harpsichord of the time, it is not distinctly for any specific type of instrument. The same is true of other compositions that have also been seen by commentators as potential clavichord music, among them the earlier of the two sonatas in A flat major (hXVI:46). Such music could be played on the clavichord, but one would need a five-octave unfretted instrument to do so effectively.
There are further reasons for doubting that these and other sonatas of the period were intended specifically for the clavichord. The relatively low string tension of many larger unfretted instruments – at least as typically encountered today – renders them less than ‘preferable’ for distinctly articulating quick Alberti basses and other lively accompaniment figures (Example 17a). Equally unidiomatic for such instruments are quick staccato octaves and parallel thirds (Example 17b), as well as the extended sequences that Haydn uses in some development passages (Example 17c).Footnote 49 Passages such as this last are described as ‘vamps’ in modern writings on Scarlatti, from whom Haydn might have derived the idea, although comparable writing occurs in all manner of eighteenth-century music; a more historical term might be perfidia.Footnote 50 On the clavichord, a player can colour such passages through dynamic shadings. Yet the overall dynamic range is limited and the sustaining power weak, by comparison with those of other types of keyboard instruments. Even on many early pianos, a narrow dynamic range and modest sustaining power make these instruments only incrementally more ‘expressive’ than a one-manual harpsichord, which remains as plausible a medium as any other type of stringed keyboard.
Haydn, (a) Sonata in D major hXVI:14/iii, bars 11–14, from Œuvres complettes, volume 12 (Leipzig: Breitkopf und Härtel, 1806); (b) Sonata in C major hXVI:21/ii, bars 22–25, from Sei sonate da clavi-cembalo (Vienna: Kurzböck, 1774); (c) Sonata in C sharp minor, hXVI:36/i, bars 51–54
A Shift in Instrumental Medium or a Change in Compositional Thinking?
The problems that modern commentators have encountered in assigning eighteenth-century keyboard music to specific types of instruments suggest that they have been asking the wrong question. Rather than shifting from one instrument to another, composers of Emanuel Bach's and even Haydn's generation may initially have thought of keyboard music as independent of any specific performing medium. As late as 1802, when Johann Nikolaus Forkel published his biography of J. S. Bach, he found it necessary to explain the composer's understanding of the difference between ‘clavier’ and ‘organ’, which was apparently not self-evident to his readers.Footnote 51 Emanuel Bach surely did understand that distinction from an early age. Yet it might not have been obvious to him before the mid- or late 1740s that his solo ‘clavier’ music was becoming more suited for performance on the clavichord than on the harpsichord. Even as he wrote the first volume of his Versuch, published in 1753, he continued to advise readers to be prepared to play pieces equally well on both instruments.Footnote 52
By the same token, only gradually, and somewhat later than elsewhere in Europe, would it have become clear to Haydn in Vienna that a player at either the clavichord or the harpsichord could realize his keyboard music only imperfectly. Dynamic markings are perhaps the most obvious although not necessarily the most important of the musical elements whose presence in a composition would have led to a new way of thinking about choice of instrument. As is well known, the number and variety of dynamic markings in his keyboard music increased over the course of Haydn's career. These included not only crescendo and diminuendo but local accents that can hardly be made audible on the harpsichord. Yet there is great inconsistency in the number and type of dynamic markings even within pieces published together (as in the 1780 set). Some works, such as the C minor sonata, might have been conceived from the outset as being more mutable in character, requiring greater dynamic flexibility. Yet even in this work, most of the original dynamic markings are of the ‘terrace’ and accentual types, calling only for local contrasts of ‘shadow and light’ and functioning like the ornaments in older music.
Composing with dynamics
Although dynamic contrasts may at first have been understood as a sort of ornamentation, not essential to the argument of a composition, dynamic effects are among the fundamental ideas of certain exceptional pieces. A sonata composed by Bach in 1758 (Wq52/6) opens softly, contrary to convention. It then grows even quieter before proceeding to an implied crescendo, as indicated by successive piano, pianissimo and fortissimo markings (Example 18). The dynamics of this opening phrase have ramifications later in the sonata, even in the second movement. Haydn may not have opened a keyboard piece in a similar manner prior to the F minor variations of 1793 (discussed below). Subsequent examples include the great Piano Trio in E flat major hXV:22, published in 1795, as well as the C major solo sonata of the same year (Example 19).Footnote 53 Yet with Haydn it is less dynamics than other features that suggest a composer whose ideas were beginning to require a more specific instrumental medium. A relatively early example involves an enharmonic modulation in the finale of the D major keyboard trio hXV:7, published by the end of 1785 (Example 20a).Footnote 54
Bach, Sonata in E minor Wq52/6/i, bars 1–4, from Zweyte Fortsetzung von Sechs Sonaten fürs Clavier (Berlin: George Ludewig Winter, 1763)
Haydn, (a) Piano Trio in E flat major hXV:22/i, bars 1–6 (strings omitted); (b) Sonata in C major hXVI:50/i, bars 1–7, from A Grand Sonata for the Piano Forte, Op. 79 (London: Caulfield, c1801)
Haydn, Piano Trio in D major hXV:7/iii, from Trois sonates pour le clavecin ou piano forte accompagnées d'un violon et violoncelle, Op. 40 (Vienna: Artaria[, 1785]) (strings omitted): (a) bars 84–96; (b) bars 1–4
The passage invites the keyboard player to give the repeated D sharps a changing colouration, reflecting the gradual realization that these notes are no longer E flats. ‘Colour’ in this case might be some combination of dynamic, articulation and timing effects, though Haydn specifies none of these. The harpsichord (‘clavecin’) is still the first instrument named in the title of the piece as published. Yet it is impossible on that instrument to make the repeated note sound any different from when the motive is first heard at the beginning of the movement (Example 20b). Elsewhere in the piece, the accompanying violin and cello can provide colour. In their silence here, however, even the most exquisite phrasing and timing by a harpsichordist cannot do justice to the enharmonic transformation of the motive, which is written into the notes of the keyboard alone at this crucial moment.
Throughout this work, as in most of Haydn's keyboard music from before the 1790s, dynamic markings remain sparse and rarely other than obvious, reinforcing what is already indicated by the texture. The Sonata in E flat major hXVI:38 opens in what seems a fairly generic keyboard idiom (Example 21a).Footnote 55 Yet the slow movement includes changing dynamics, in a passage whose busy, low left-hand part is not easily subordinated to the melody except on a clavichord or piano (Example 21b). The octaves that accompany the main theme of the final movement are less unsuited for the harpsichord, but this type of writing, which requires the hand to leap quickly, is hardly idiomatic for the clavichord (Example 21c). That leaves some sort of piano as the most satisfactory instrumental medium.
Haydn, Sonata in E flat major hXVI:38, from Sei sonate per il clavicembalo, o forte piano: (a) first movement, bars 1–4; (b) second movement, bars 92–131; (c) third movement, bars 1–4
Given its relatively early date, hXVI:38 is unlikely to have been written specifically for piano. Rather, like the trio hXV:7, it presents musical ideas that are most fully realizable on a piano – especially the larger grand pianos of the late eighteenth century. For discussions of dynamic effects, articulation and sonorities we lack an analytical vocabulary that is as well developed as that used for themes and motives. Yet compositional ideas involving dynamics or ‘colour’ may be equally crucial in music of the later eighteenth century, as with the initial piano and subsequent crescendo in Bach's E minor sonata. Occurring at the outset of the composition, these markings indicated that dynamic variability would be one of the piece's basic ideas. This dynamic changeability continues in the slow second movement, where forte phrases alternate with piano echoes (Example 22).Footnote 56
Bach, Sonata in E minor Wq52/6/ii, bars 1–7
This sonata was published well before the appearance of anything similar by Haydn.Footnote 57 But when Haydn does open the late C major sonata (hXVI:50) in a comparable way, the initial piano marking has a ramification in the famous pianissimo passages heard in the development and recapitulation sections. These are marked ‘open pedal’, indicating performance without dampers (Example 23). The marking is unique in Haydn's keyboard music, but these are not the only passages in his late keyboard compositions that would have been inconceivable apart from one of the newer types of fortepiano. The theme of the F minor variations (hXVII:6) opens presumably piano and then crescendo (Example 24).Footnote 58 What rules out not just the harpsichord but also, probably, the clavichord, is the requirement to sustain the pensively moving lower voices – not only the suspensions in bars 4–6 and 8–11 but the uniquely Haydnesque chromaticism that accompanies the latter.
Haydn, Sonata in C major hXVI:50/i: (a) bars 73–76; (b) bars 120–124
Haydn, Variations in F minor, hXVII:6, bars 1–12, from autograph manuscript, New York Public Library, JOE 72-13
For Haydn, the critical changes in compositional thinking may not have taken place before the mid-1780s. This was the period during which Bach was publishing rondos explicitly for the piano, but by then he had been adapting to dynamic keyboard instruments for several decades. That Haydn apparently paid less attention to keyboard idiom might have reflected the late arrival of the piano in Vienna.Footnote 59 It could also be attributed to his not having been primarily a keyboard soloist. If he indeed spent much of his performance time leading the Eszterháza ensemble from the violin,Footnote 60 he would not have focused his substantial creative powers on the invention of new keyboard idioms, at least not to the degree that Mozart or the Bachs did. Perhaps, in addition, Haydn simply lacked interest in new instruments and instrumental idioms, an attitude that seems to have been characteristic of mid-century Vienna. This contrasted with Berlin, where new types of flute as well as keyboard instruments were adopted at the royal court during the 1740s, influencing fashionable music-lovers across the northern parts of the empire.Footnote 61 The piano idiom that Haydn eventually achieved is utterly different from that of Emanuel Bach, just as the pianos that he knew were very distinct instruments. Both, however, came to recognize that one could compose specifically for the piano, not merely for generic keyboard – just as one might compose for string quartet, as opposed to a string ensemble of unspecified size with continuo.
Haydn and Bach
Haydn famously acknowledged a debt to Bach, especially his ‘first six sonatas’.Footnote 62 This report has generated considerable discussion, but Haydn also at one point asked the publisher Artaria to send him Bach's two last keyboard works, plausibly identified by Wolfgang Fuhrmann as the final instalments of the series Für Kenner und Liebhaber.Footnote 63 Published in 1785 and 1787 respectively, those two volumes (Wq59 and Wq61) included four of the composer's unique modulating rondos. The latter likely provided a model for Haydn's Fantasia in C major (hXVII:4), as seems clear not only from the varied restatements of the main theme in various keys but also the use of modulating passages based on arpeggiation – here interrupted several times by mysterious fermatas (Example 25; compare Example 1d).Footnote 64 Much of the passagework is eventually recapitulated, and Haydn's fantasia concludes with a written-out cadenza. These are all features that Haydn would have found in Bach's rondos, especially the great one in G major (Wq59/2) from the penultimate collection of Kenner und Liebhaber.Footnote 65 Another work by Bach, the C major fantasia from the final volume in the series (Wq61/6), also comes into question.Footnote 66 The latter resembles the composer's rondos more closely than it does his other fantasias, in this respect bearing comparison with Haydn's fantasia in the same key.
Haydn, Fantasia in C major, hXVII:4, from Fantasia per il clavicembalo o Forte-Piano, Op. 58 (Vienna: Artaria, c1789): (a) bars 1–4; (b) bars 164–167; (c) bars 184–192
C. P. E. Bach's ‘last’ keyboard works
In the case of Haydn's Fantasia in C major, the claim for Emanuel Bach's influence is plausible because one can draw direct parallels between specific compositions, and Haydn had readily documentable access to the postulated models. The subscriber lists printed within Bach's two publications show that multiple copies were sent to both Artaria and ‘Baron von Swieten’ in Vienna, the latter being doubtless the musical patron and collector Gottfried van Swieten.Footnote 67 Bach's rondos would have attracted attention not only for their singular form but also for their designation as piano pieces. Haydn would not have missed that, despite seeing his own pieces still published with the commercially advantageous alternative assignment to the harpsichord. Yet although Haydn's fantasia is surely a piano piece, his piano idiom is distinct from Bach's – on the whole somewhat heavier by this date (c1790) and incorporating the type of motoric accompaniment figures avoided by Bach (see, for instance, Examples 25b and c). Nevertheless, the texture of Haydn's fantasia is somewhat lighter than in the composer's last sonatas and piano trios. This might reflect his composing the work while still having in mind the types of instruments favoured at Vienna, rather than the English pianos with which he became acquainted a few years later.Footnote 68 Regardless of the specific type of instrument, however, this is a distinctly pianistic composition as opposed to one conceived generically or for the harpsichord.
Haydn's G major Capriccio (hXVII:1) appears to present a comparable case, for, like the C major fantasia, it would seem to have been a response to Emanuel Bach's modulating rondos of the 1780s – if only it were not dated 1765 in Haydn's autograph score. Yet the rondo-like form of the Capriccio, as well as its combination of the serious with the burlesque, allies it with an earlier composition of Emanuel Bach. Bearing the unexplained title L'Aly Rupalich, this famously perplexing character piece, composed in 1755, came out in an anthology during 1762 or 1763 (Example 26).Footnote 69 On the surface, the piece has little to do with Haydn's capriccio. Yet seeing something as unbuttoned as Bach's piece, by a respected composer, would not have discouraged the still youthful Haydn from writing something equally outré.
(a) Bach, ‘L'Aly Rupalich’, Wq117/27, bars 1–7, from Musikalisches Mancherley (Berlin: George Ludewig Winter, 1762); (b) Haydn, Capriccio in G major, hXVII:1, bars 1–5
C. P. E. Bach's ‘first’ six sonatas
As influential as any of Emanuel Bach's later keyboard works might have been on Haydn, scholars have been more exercised over the question of what Haydn meant by Emanuel Bach's ‘first’ six sonatas – not least because these might have exerted a formative influence on the younger musician. Both composers dedicated their first publications of keyboard music to their employers; Haydn might have noted that Bach's ‘Prussian’ Sonatas were dedicated to Frederick ‘the Great’, just as his own collection of 1774 was presented to Nikolaus Esterházy. Yet the identification of the Bach set as the one meant by Haydn is by no means certain. Seeking a more concrete basis for an identification, Ulrich Leisinger has considered which of Bach's keyboard sonatas actually circulated in Vienna during Haydn's time.Footnote 70 Others have sought internal evidence, chiefly in the form of thematic parallels, which are easy enough to find. Yet the presence of common motivic ideas may be less significant compositionally than elements of the music that lie beneath the surface.
For instance, the opening of Bach's Sonata in C minor of 1741 (Wq48/4) has been found to share ‘some common features’ (‘einige gemeinsame Züge’) with that of Haydn's early Sonata in G major hXVI:6.Footnote 71 Yet these features seem to be limited to vaguely similar rhythm and phrasing (Example 27a and 27b). A more meaningful parallelism might be drawn to the opening of Haydn's C minor sonata, which shares with Bach's sonata not only its key but the insistence on appoggiaturas decorating the notes G and B♮.Footnote 72 Whether the fermatas on dominant chords that interrupt both movements also constitute significant parallelisms is debatable. In any case, correspondences of these types could demonstrate only that Haydn was thinking compositionally along lines similar to Bach's, not that he was actually imitating the earlier piece.
(a) Bach, ‘Prussian’ Sonata No. 4 in C minor, Wq48/4/i, bars 1–6, from Sei sonate per cembalo (Nuremberg: Balthasar Schmid, 1742); (b) Haydn, Sonata in G major hXVI:6/i, bars 1–4, from Œuvres complettes, volume 12 (Leipzig: Breitkopf und Härtel, 1806)
Leisinger, however, showed that another relatively early sonata movement by Haydn – the Moderato of hXVI:18 – proceeds for some twenty-four bars, nearly to the end of the exposition, in a way that parallels a rather different work of Bach's (Example 28). The publication containing the latter (Wq50/5) was actually his fourth printed set of keyboard sonatas, the Reprisen-Sonaten, in which repeated passages are furnished with written-out variations.Footnote 73 Melodic and rhythmic parallels between the two sonata movements are indeed striking, yet are there any deeper relationships? Haydn's movement incorporates no substantial variations, although neither does Bach's, despite its inclusion in one of the Reprisen-Sonaten. On the other hand, the Bach example is one of very few quick sonata movements by the composer that avoids some version of binary or sonata-allegro form. It is instead a sort of rondo, albeit one with sonata-like elements. Precisely for that reason, Bach's composition might have attracted the attention of a thoughtful musician interested in musical form or design. If Haydn did come across it in the years around 1760, surely he would have recognized its unusual qualities and studied it carefully.
(a) Bach, Sonata in B flat major Wq50/5/iii, bars 1–11, from VI. Sonates pour le clavecin avec des reprises variées; (b) Haydn, Sonata in B flat major hXVI:18/ii, bars 1–10, from Œuvres complettes, volume 1 (Leipzig: Breitkopf und Härtel, 1799), based on Ulrich Leisinger, Joseph Haydn und die Entwicklung des klassischen Klavierstils bis ca. 1785, Example 7.3 (Laaber: Laaber, 1994), 287
Influence and echoes
Speculation of this sort can only suggest how one creative musician might have wound up, perhaps unconsciously, closely echoing the work of another. It cannot settle the issue of which sonatas by Emanuel Bach were actually known to Haydn, let alone influenced him. At best, one can conclude, as Elaine Sisman put it, that a rondo with variations also from the Reprisen-Sonaten – the concluding Sonata in C minor Wq50/6 (in a single movement) – is a ‘more than merely plausible’ model for Haydn's examples of the similar form known as ‘alternating variations’.Footnote 74 Among the latter is the final movement of Haydn's G minor sonata, hXVI:44, one of the composer's ‘expressive’ sonatas of the mid-1760s. As such, the sonata might, incidentally, seem particularly suitable to the clavichord, but at least one passage would be awkward on any but a large unfretted instrument (Example 29).
Haydn, Sonata in G minor, hXVI:44/i, bars 9–12, from Three Sonatas for the Pianoforte or Harpsichord, Op. 53 (London: Longman & Broderip, c1788)
Be that as it may, the continuation of this particular sonata has been seen as deliberately emanuel-bachisch, inviting comparison with a work such as Wq50/4, which at one point employs equally bizarre figuration (compare Examples 30a and 30b). Passages such as this might even have been what led a contemporary English observer to claim that Haydn's first two printed sets of sonatas were not influenced by, but rather were meant ‘to ridicule Bach of Hamburgh’ by way of parody.Footnote 75 That idea has been refuted,Footnote 76 but whether parodistic or merely playful, Haydn here displays a sense of humour shared with the older composer. Still, even if Bach's wit provided a ‘more than merely plausible’ model for Haydn's, that is too vague to serve as an argument for direct influence.
(a) Haydn, Sonata in G minor, hXVI:44/i, bars 13–14; (b) Bach, Sonata in D minor Wq50/4/i, bars 88–91, from VI. Sonates pour le clavecin avec des reprises variées
If Haydn did pick up the idea of alternating variations from Emanuel Bach, he uses it in a very different way. Bach's C minor sonata, following a tradition that went back at least to Corelli's Follia variations (Op. 5 No. 12), served as a concluding tour de force for the volume in which it was published. Yet for Bach those variations remained a demonstration of what was in principle a performance practice, a type of improvisatory embellishment.Footnote 77 Haydn, on the other hand, uses the principle of embellishment structurally, incorporating alternating variations into a distinctly classical design: he returns to the opening theme in its original form, unvaried, to serve as a ‘reprise’ (Sisman's term) just before the coda of the F minor variations – a structural function absent from most of Bach's variation works.
The question of which of Bach's pieces actually influenced Haydn not only seems unanswerable but is probably the wrong one to ask. To pose such a question is really to ask how one creative mind responds to the work of another. Yet an inventive composer might be profoundly moved – set on a new path – by ideas that would not seem particularly significant to anyone else. Harrison noted a new ‘orthography’ in Haydn's keyboard music of the later 1760s, following his encounter with Emanuel Bach's publications of a few years earlier.Footnote 78 Seeing Bach's highly rationalistic systems not only of ornament signs but of figured-bass symbols, Haydn might have become aware of a certain negligence in his own notational practice, resolving thenceforth to indicate his intentions more precisely. Reading Bach on the need for varied reprises, then studying written-out examples in the Reprisen-Sonaten, Haydn might have considered how to incorporate such decoration into the deeper structure of a composition. Examining pieces that depart from conventional musical forms, even if only for descriptive or humorous purposes, Haydn might have been inspired to rethink formal design at a fundamental level.
What could keyboard music be?
Even more basically, Bach could have given Haydn a broadened view of what was possible when composing for keyboard instruments. The Italianate tradition of keyboard music that Haydn encountered in Vienna in his youth focused on two sharply different but equally generic types of writing: strict, quasi-vocal polyphony in a tradition that went back to Frescobaldi and Froberger, and homophonic pieces in a simple popular style. Examples of the former include the canzonas and ricercars attributed to Georg von Reutter (or his son of the same name, who was Haydn's teacher), and, of the latter, the variously titled divertimentos and similar compositions for solo and accompanied keyboard by Wagenseil and the Monn brothers. One of the most admired composers of the time was Hasse, who wrote, besides some striking operas and oratorios, many popular but facile keyboard sonatas and concertos. If Haydn knew those pieces, they would have demonstrated to him that an admired composer of sacred and dramatic music could coast when writing for instruments alone.
Curiously, Hasse may have been the one older contemporary known personally to both Haydn and Emanuel Bach, serving each as a professional model, if not exactly a mentor. Hasse praised works by both younger composers, referring to one of these as the best symphony he had ever seen. That was (perhaps surprisingly) Emanuel Bach's E minor work of 1756 (Wq178), which demonstrated that a symphony could be more than a noisy opening for a night at the theatre.Footnote 79 Bach's E minor symphony survives both as an orchestral work and in a keyboard version. Although the composer's responsibility for the latter is not entirely assured, it illustrates a symphonic type of writing for the keyboard also found in Bach's sonatas of the 1750s and later.Footnote 80 The E minor sonata shown in Examples 18 and 22 is another instance of this type; a later one is the Sonata in B flat major from the penultimate collection for Kenner und Liebhaber (Wq59/3). Many of Mozart's solo sonatas show elements of this style, which can, however, sound both pretentious and derivative, not entirely suited to keyboard instruments. Echoes of it might be heard in Haydn's late Sonata in E flat major hXVI:52. Yet that work makes only sparing references to actual orchestral style, and the sonata remains more idiomatic to the keyboard than the more literal evocations of actual symphonies by both Bach, in one of his last printed keyboard sonatas, Wq59/5, and Mozart, in his ‘Dürnitz’ sonata k284/205b (Example 31).Footnote 81
(a) Haydn, Sonata in E flat major hXVI:52/i, bars 38–39, from autograph manuscript, Washington, Library of Congress, ML96 .H364 Case SSF (1794); (b) Bach, Sonata in B flat major Wq59/5/i, bars 93–112, from Clavier-Sonaten und Freye Fantasien nebst einigen Rondos fürs Fortepiano für Kenner und Liebhaber, Wq59 (Leipzig: author, 1785); (c) Mozart, Sonata in D major k284 (205b)/i, bars 9–13, from autograph manuscript, Krakow, Biblioteka Jagiellońska, Mus. ms. autogr. W. A. Mozart 279–284, 330, 455
Haydn thus avoided a keyboard idiom found in innumerable works of the time, including those of his best contemporaries. Even Bach's ‘symphonic’ sonatas, however, demonstrated that keyboard music could be more than either a pleasant divertissement or an exercise in strict counterpoint. Keyboard pieces could be intellectually stimulating in ways that did not involve counterpoint, and they could be challenging not only technically but emotionally. Among the challenges posed by Emanuel Bach's keyboard compositions was the performative one of projecting emotions or affects to listeners while playing wordless music on a stringed keyboard instrument. Contemporary accounts, especially the famous one by Burney, depict Bach's playing at his Silbermann clavichord as a reflection of the composer's personal emotional state. Burney described Bach as inspired, using a word which at the time probably implied something more visceral than it does today. ‘Inspiration’ invoked the classical image of an ancient Greek sibyl, literally inspiring or breathing in divine vapours from a mysterious sacred source.Footnote 82 Late in life, when Haydn described his working method, he wrote as if his compositions likewise began as a manifestation of his own psychological state, as he sat at his clavichord and improvised.Footnote 83
Emanuel Bach did not record similar thoughts. Yet the presence of the clavichord in these accounts of both composers reflects its association in pre-romantic German-speaking Europe with private, sensitive musical performance. Another trope in writing of the time is the paradox that great music might be written on a tiny instrument limited in both compass and dynamic range. Haydn's account probably alluded to this, as Emanuel Bach certainly did when he informed Forkel that two sonatas resembling fantasias had been composed on a travelling clavichord with a short octave.Footnote 84 In neither case was the composer indicating that his music was written specifically for the clavichord in question, which in Bach's case could not even have provided all the necessary notes. Yet each associates keyboard music with an instrument known for its utility for subjective expression, not for either learned counterpoint or light entertainment.
That the piano eventually took the place of the clavichord as the most ‘expressive’ keyboard instrument was due not only to its broader dynamic range but to its greater sustaining power and its utility in all keys. In its larger forms it was also better equipped for the types of public performances by virtuosos that were coming into vogue as the nineteenth century approached. That these aspects of the piano came to be viewed as desirable was reflected in changes in instruments and in musical style that took place during the second half of the eighteenth century – and these changes came with a price. The sustaining power of later pianos, together with keyboard actions and techniques that favoured legato over articulate performance, tended to smooth over details such as the short slurs and frequent ornaments that complicate or problematize motion from one note to the next in late baroque and early classical music. Ornaments and minute articulations are more readily controlled on the actions of earlier types of stringed keyboard instruments. Although composers and writers in northern Europe, including Bach and Daniel Gottlob Türk, continued to focus on details of ornament and articulation to the end of the century, by the 1780s such things had become less important for other musicians, Haydn among them.Footnote 85 His shift to writing distinctly for the piano was an element in his achievement of what used to be called the ‘mature classical style’.Footnote 86 We might tentatively associate it with the end of Haydn's deep involvement with private court opera and his renewed engagement with the public after 1780 through printed instrumental works.Footnote 87
Haydn scholars now see these developments as involving what they describe as the rhetoric of his music.Footnote 88 Indeed, the metaphorical rhetoric of Haydn's later compositions, including those for keyboard, involves speaking in longer paragraphs that are more evenly flowing, less subdivided or articulated into smaller units, than in earlier works. The piano, by virtue of its capabilities for extended crescendi and diminuendi, as well as for long legato lines, is better able to project or convey this rhetoric than the clavichord or the harpsichord. The new approach to instrumental musical rhetoric helps explain why one might feel that by the late 1780s Haydn was doing what neither he nor Bach had done in earlier keyboard music: shifting toward a distinctively pianistic idiom, albeit only gradually and somewhat imprecisely at first. The piano idiom in Bach's late rondos remains only subtly different from his writing for generalized ‘clavier’ in other compositions. Haydn's piano writing is ultimately more distinct, not only from his own earlier keyboard writing, but also from that of his younger contemporaries, including J. C. Bach and Mozart: less focused on new types of virtuoso figuration and accompaniment patterns, and more on characteristic types of keyboard texture and motivic development (as seen, for instance, in Examples 17, 20 and 25). The best player of either harpsichord or clavichord must feel inadequate when facing the late keyboard works of both composers, especially Haydn's final sonata in E flat and the London piano trios. Nevertheless, in almost everything leading up to them, the piano adds only a layer of refinement to music that remains expressible, in large part, on older types of instrument.
