FROM ‘GALANT STYLE’ TOWARDS PRAGMATIC THEORIES OF MUSICAL LANGUAGES

‘Our language can be seen as an ancient city’, writes Ludwig Wittgenstein, ‘a maze of little streets and squares, of old and new houses, and of houses with additions from various periods; and this surrounded by a multitude of new boroughs with straight regular streets and uniform houses’ (Ludwig Wittgenstein, Philosophical Investigations (Oxford: Blackwell, 1953), §18). Wittgenstein's point is that human language is not a coherent system based on a single principle, but a set of conventions that has proliferated continuously over centuries. And the same is true for musical language. In the late eighteenth-century musical language of Haydn and Mozart, for example, we see elements as old as three centuries (parallel chords, series of 7-6 suspensions and the old ‘Romanesca’) next to elements of two centuries’ age (most cadences and the ascending-fifths sequence, for example) and those only several decades old (such as ‘galant’ dissonance treatment). One of the lessons we have learned is that a rich florilegium or zibaldone of musical commonplaces was (and is) a better resource for an aspiring composer-improviser in this language than a set of abstract principles.

But it is not only a question of galant or Italianate music. Both the subtly humorous tone and focus on galant music in Gjerdingen's book have made it too easy a target for its critics. David Damschroder, for example, has rejected Gjerdingen's approach mainly because it ‘seeks to place Haydn and Mozart within the mainstream of Italianate practice’ (Harmony in Haydn and Mozart (Cambridge: Cambridge University Press, 2012), 168). Damschroder takes Leopold Mozart's dismissive words about Italian music and musicians to ‘validate’ the use of ‘northern analytical techniques’, which he sees as the root of his own analytical perspective based largely on Schenkerian analysis and Roman numerals. While this Italy–Germany dichotomy has been put back on the table by ‘partimentisti’ themselves, Gjerdingen's findings and the potential of partimenti cannot be dismissed so easily.

In much galant music schemata, such as those so meritoriously collected by Gjerdingen, are present in an especially clear and straightforward way. But the time is ripe to extend our perspective. First, to other eighteenth-century styles without, however, falling into the circular trap of identifying all compositions with allegedly galant schemata as ‘galant’ (or as remnants of a ‘courtly art’), and without setting up a dichotomy between conventional galant music based on schemata and the works of heroic composers who consistently transcend these conventions. Second, the schema or Satzmodell approach needs to be more thoroughly carried beyond the boundaries of the eighteenth century. The approach, mutatis mutandis, is as revealing for late fifteenth- and sixteenth-century music as it is for nineteenth-century music (see my ‘Romantic Thoroughbass: Music Theory between Improvisation, Composition, and Performance’, Theoria 18 (2011), 5–36, and my ‘“Aut propter Devotionem, Aut propter Sonoritatem”: Compositional Design of Late Fifteenth-Century Motets in Perspective’, Journal of the Alamire Foundation 9/1 (2017), forthcoming). In working with schemata, we should proceed from questions of one specific (galant) style towards pragmatic theories of musical styles understood in the sense of Wittgenstein's human language.

FROM ‘PARTIMENTO’ TO ‘GENERALBASS’

The eclectic nature of musical language in its multiple historical layers suggests we should view big theories with some scepticism, especially when they are based on supposedly timeless principles. On the other hand, this does not mean there is no need for theorizing (or at least conceptualizing) different elements of the ‘ancient city’ that is musical language. In the eighteenth century there were many attempts to engage with the different layers of musical language. The flexible conceptual framework for all of them was thoroughbass: the theory and practice of reading a polyphonic fabric from a figured or unfigured bass. And this brings me to my next point.

For a long time, a green salad vegetable called ‘Rauke’ lived in the shadows of German cuisine (owing to its tangy and bitter taste) until it finally found culinary life under a new identity, conquering hearts and mouths virtually everywhere under the Italian name ‘rucola’. Something comparable could be said about the term ‘partimento’. Certain concepts, I have the impression, have re-entered the stage under this new name, concepts that used to be known under the banner term Generalbass or ‘thoroughbass’. Certainly, compared to the disarming Mediterranean nonchalance of ‘partimento’, the words ‘Generalbass’ or ‘thoroughbass’ can sound dusty, pompous, clumsy. But for me thoroughbass reveals the full – but largely untapped – potential of partimenti. A renaissance of thoroughbass as the true foundation of eighteenth-century performance, improvisation, composition and analysis would be an even more significant event than the ‘discovery’ of schemata. Discussions of galant schemata that continue to use Schenkerian analysis and Roman numerals as theoretical frameworks only change the foreground, as it were, whereas a true integration of thoroughbass would profoundly change the conceptual, theoretical and practical background for a new analytic approach to eighteenth-century music.

How could thoroughbass have been so neglected in discussions of schemata? While in eighteenth- and nineteenth-century thought the discipline of thoroughbass was seen as forging a unity between practice and theory, comprising the complete grammar of musical language, the twenty-first-century revival of ‘partimenti’ has focused almost exclusively on the practical side. Much emphasis has been put on the implicit, embodied, practical, non-literary, non-theoretical knowledge of eighteenth-century musicians, and rightly so. The partimento movement has seemed like a collective gasp of relief revealing much about the weighty theories long impressed on scholars and students without much alternative. But we surely need a course correction here if we do not want to throw out the baby with the bathwater. I am reminded of a discussion about medieval music theory between Edward Lowinsky and Richard Hoppin in the 1940s and 1950s. While Lowinsky concluded that ‘the medieval composer was eminently practical’, Hoppin insisted that ‘he was more often eminently theoretical’ (Edward Lowinsky, ‘The Function of Conflicting Signatures in Early Polyphonic Music’, The Musical Quarterly 31/2 (1945), 241; Richard Hoppin, ‘Partial Signatures and Musica Ficta in Some Early 15th-Century Sources’, Journal of the American Musicological Society 6/3 (1953), 204). In a sense, of course, both were right. And the same could be said about the eighteenth century. The rule of the octave, for example, was on the one hand an eminently pragmatic rule of thumb for realizing unfigured basses. On the other hand, it was subject to much speculation and, as Ludwig Holtmeier has shown, a matrix for concepts of tonality.

In a word: the time is ripe for carefully re-theorizing eighteenth-century music, for turning to ‘thoroughbass as music theory’ (see Thomas Christensen's article of that title in Partimento and Continuo Playing in Theory and in Practice, ed. Dirk Moelants (Leuven: Leuven University Press, 2010), 9–41). To do so, we need not fall back into what Christensen refers to as ‘monumental’ theories. The theories to start with now are the small theories of thoroughbass and accompagnement, some of which should subsequently change our perspective on the ‘big’ theories that frequently grew out of the ‘small’, as Christensen observes. An obvious example is Heinichen's monumental Der Generalbass in der Komposition of 1728, discussed by Holtmeier (‘Concepts of Tonality’) and Christensen (‘Monumental Theory’) as a thousand-page attempt to codify thoroughbass as Music Theory. There are worlds to discover here. Books like Heinichen's contain endlessly nuanced and detailed discussions of chords, progressions, cadences, dissonances and so on – hence, a complete vocabulary for a musical language based on both the contrapuntal heritage and on the modern practice of keyboard accompaniment and improvisation. To re-theorize partimenti and eighteenth-century music means to move beyond a pioneer phase characterized by scholars ‘discovering’ more and more schemata on which they plant their flags by giving them idiosyncratic and personal names. It means to proceed from a phase of reification of single schemata (something certainly necessary, especially for pedagogy) to a phase in which the entirety of tonal schemata can be seen against a rich eighteenth-century theoretical background.

How closely theoretical reflection and practical application stand together in the context of eighteenth-century theories may be illustrated by the following simple example. Mozart's reworkings of other composers’ keyboard sonatas into concertos afford us a valuable opportunity for gazing into an eighteenth-century composer's workshop and observing the path from sonata form to concerto-sonata form. What concerns us here are some contrapuntal details in Mozart's arrangement. Consider bars 26–29 of Johann Christian Bach's sonata Op. 5 No. 2 (Example 1) and the accompanying string parts Mozart added for his concerto arrangement (Example 2). Bach's two-part piano texture looks purely consonant, but Mozart added a dissonant accompaniment. What Mozart exploits here is a crucial aspect of thin, frequently two-part and consonant ‘galant’ settings. While it is well known to modern theory that dissonant passages can be reduced to a consonant background, the opposite is also true: consonant passages, as eighteenth-century theory teaches explicitly, frequently imply dissonances that can be heard without being present. Rameau's concept of ‘sous-entendu’ (notes that are implied without being actually present), frequently ridiculed as being eminently theoretical, is in fact something eminently practical (see Ludwig Holtmeier's discussion of Handel's thoroughbass exercises in Georg Friedrich Händels Aufzeichnungen zur Kompositionslehre, ed. Ludwig Holtmeier, Johannes Menke and Felix Diergarten (Wilhelmshaven: Noetzel, 2013), 51–242).

Example 1 Johann Christian Bach, Sonata D major Op. 5 No. 2/i, bars 26–29, transcribed from Six Sonatas for the Piano Forte or Harpsichord (London: Welcker, 1765) <www.imslp.org>

Example 2 W. A. Mozart, Keyboard Concerto k107 I/i, bars 54–57, transcribed from Wolfgang Amadeus Mozart, Neue Ausgabe sämtliche Werke, series 10, volume 28/2, Bearbeitungen von Werken verschiedener Komponisten: Klavierkonzerte und Kadenzen, ed. Walter Gerstenberg and Eduard Reeser (Kassel: Bärenreiter, 1964) <dme.mozarteum.at/DME/nma>

Bach's two-part texture implies three parts: between the lowest part (a pedal on e¹) and the melody on top there is a middle voice that uses one of the most common eighteenth-century sequences (bars 27–28, moving up a second and falling a third) and is the actual leading part governing the progression (the guide). The typical – that is, dissonant – realization of this progression would be a chord on the ascending notes (Example 3a). The historical, contrapuntal origin of this progression is actually a series of suspensions in the upper parts, to which the bass forms an additional, supplementary part (see, for example, how Angelo Berardi derives this progression in his Documenti armonici (Bologna: Monti, 1687), 154). Together with the bass, the upper parts display the quinta dissonans, a fifth made dissonant (and thus always resolving down a step) by the sixth above it, respectively forming the parte patiente and parte agente in a syncopation (to borrow the terminology of Giovanni Artusi and Andreas Werckmeister).

Example 3 Contrapuntal analysis of Bach's melody

What we see in Bach's seemingly ‘uncontrapuntal’ and consonant two-part setting, then, is a keyboard figuration of a dissonant four-part texture (see Examples 3b to 3d, adding the pedal on e¹ as the lowest part). His melody in bar 27 consists of the sixth, f♯², rendering dissonant the e² virtually sustained from bar 26 (a phenomenon referred to as tenue or heterolepsis) and then, leaping back into the original part, the resolution of e² to d³. Mozart had probably immediately seen, heard and felt in his fingers that Example 3a resides virtually within Example 3d. What he did in his accompanying string parts, then, was to recreate the dissonant and polyphonic setting: the dissonances sous-entendues became actual dissonances. Incidentally, the resulting chains of suspensions are highly typical of the dissonant sequences in his later piano concertos.

We often hear the presumptuous claim that modern theories of harmony are superior to eighteenth-century theories because the latter exhibit only a rudimentary understanding of harmonic progressions. Whoever seriously and patiently explores eighteenth-century thoroughbass manuals and the theories in their orbit soon realizes that the opposite is true. These books offer an endlessly rich and nuanced vocabulary of harmonic and contrapuntal phenomena of all sorts. Tragically, their complexity was the reason for their downfall: partimento practice and related theories ‘could raise scant opposition to the manifest logic of Rameau's principle of inversion’ (Holtmeier, ‘Concepts of Tonality’, 42). The tempting logic of ‘modern harmony’, the demand for ‘simplification, systematisation, and rationalisation’ (Christensen, ‘Monumental Theory’, 210) rolled over the artisanal theories and practices of thoroughbass. Today, however, we simply cannot afford to ignore them.

BEYOND ‘HARMONY’ AND ‘COUNTERPOINT’

Anton Reicha, a friend of Beethoven and Haydn, wrote in the early nineteenth century that ‘the terms counterpoint and harmony are synonymous’ (Vollständiges Lehrbuch der Composition, ed. Carl Czerny (Vienna: Diabelli, 1832), 711). It is a truism in modern textbooks that eighteenth-century counterpoint is based on harmony. But now that our understanding of eighteenth-century harmony has changed profoundly through partimenti and thoroughbass, this truism finally starts to make sense because it turns out that the opposite is true: eighteenth-century harmony is based on counterpoint. In simple terms, a polyphonic fabric moving above a real existing bass line always acts as a counterpoint against this bass. ‘To play thoroughbass means to produce proper counterpoint on a keyboard’, wrote Friedrich Erhard Niedt in 1717 (Musicalische Handleitung, part 3 (Hamburg: Schillers Erben, 1717), 2). Thoroughbass looks like harmony when moving in first-species counterpoint against the bass, but it starts to look like counterpoint when moving in florid (or even imitative) style. The difference between harmony and counterpoint, put plainly, is a question of rhythm. There is probably no source that is so telling in this regard as Luigi Cherubini's pedagogical treatise Marches d'harmonie (Paris: Heugel, no date). Cherubini gives a catalogue of virtually all imaginable sequential progressions in the bass, starting from a simple figured bass and ending with complex contrapuntal and imitative fabrics over the same bass. It wouldn't make sense against such a backdrop to distinguish between counterpoint and harmony in the Bach/Mozart example discussed above.

The lesson to learn from all this is that we need to rethink how we teach compositional techniques of the so-called ‘common-practice period’; we need especially to rethink the curricular separation of counterpoint and harmony. In addition, more attention should be paid to those countless eighteenth-century stile antico pieces and to the vast spectrum of compromises between ‘strict’ and ‘free’ styles, a binary opposition that needs to be deconstructed. This is not only a question of theory, analysis and pedagogy, for it leads us to revisit a historical topic that has been filed away for some decades now: the integration of contrapuntal or ‘learned’ elements into the galant style, which was an explicit concern in eighteenth-century compositional instruction (see, for example, Johann Friedrich Daube's Der musikalische Dilettant (Vienna: Edler von Trattnern, 1773)). This topic long dominated discussions of the ‘classical style’ in general and of Haydn's music in particular, until James Webster revealed the ideological and historiographical problems of a narrative that considers ‘classical style’ a synthesis of ‘binary oppositions’ such as homophony vs polyphony (Haydn's ‘Farewell’ Symphony and the Idea of Classical Style: Through-Composition and Cyclic Integration in His Instrumental Music (Cambridge: Cambridge University Press, 1991), 342). But it does need a fresh, bottom-up perspective, one that might take its point of departure from the explicit discussions in eighteenth-century composition manuals.

TOWARDS A HISTORICALLY INSPIRED FORMENLEHRE

When music theorists disagree about (or diagnose ambiguities in) what seem to be rather straightforward pieces, their discussions typically reveal more about the analytical methods employed than about the pieces themselves. This is especially true for ‘ambiguities’ of musical form, and a striking example recently revolved around the slow movement of Haydn's ‘Farewell’ Symphony. In his Harmony in Haydn and Mozart (155–156) David Damschroder challenged an analysis of that movement in Webster's classic study (Haydn's ‘Farewell’ Symphony and the Idea of Classical Style, 57). While Webster saw the onset of the ‘transition’ in bar 21 of that movement, Damschroder pulled out all the stops of his analytical method to demonstrate that fixing that moment in bar 17 is superior to Webster's reading.

Disregarding all historical accounts of form, neither author questions whether something in that movement must be called a ‘transition’ at all. In fact, Haydn's movement up to the repeat sign follows what Heinrich Christoph Koch describes as the usual form for this part (using examples from Haydn's symphonies to illustrate his point): a cadence in the main key right at the beginning, a phrase ending with a half close in the main key, another phrase that ends with a half close in the secondary key and a long series of phrases trying to find (and then reiterating) a cadence in the secondary key, this last coloured, in the ‘Farewell’ movement, by turns to the minor that are so typical of Haydn's early symphonies. In Koch's discussion, the term ‘transition’ is neither present nor necessary: the phrase leading to the half close in the main key is directly followed by the phrase leading to a half close in the secondary key, and no further functional distinction between these two phrases is made. I have argued elsewhere that Koch's relationship to Haydn and the classical style should be treated with greater care than is typical, that Koch certainly cannot be read as a universal Formenlehre of classical style, and that the perspective needs to be broadened from Koch at least to early nineteenth-century accounts of form such as those of Reicha and Czerny (Felix Diergarten, ‘“Auch Homere schlafen bisweilen”: Heinrich Christoph Kochs Polemik gegen Joseph Haydn’, Haydn-Studien 10 (2010), 78–92). I would argue that Koch's ‘punctuation form’ is an indispensable tool for correcting later Formenlehre’s fixation on the functions of single segments in a piece at the expense of a traditional concept of cadential articulation.

The pluralistic but historically inspired Formenlehre we need has existed for some time. It took hold in German-language music theory during the early 1980s, based on Carl Dahlhaus's and Wolfgang Budday's work on punctuation form as described by Koch and Riepel (see, for example, Budday, Grundlagen musikalischer Formen der Wiener Klassik (Kassel: Bärenreiter, 1983)). But it has remained a subculture in Anglo-American theory for several reasons. First, Schenkerian theory relegated Formenlehre in the narrow sense to mere sideshow status. Second, where a historical perspective on musical form was taken, it generally followed Charles Rosen in focusing on the roots of sonata forms in compositional practice, that is, in baroque dance movements, aria forms and so on, but largely neglecting eighteenth-century writings on form. Finally, the highly influential textbooks of William Caplin (Classical Form: A Theory of Formal Functions for the Instrumental Music of Haydn, Mozart, and Beethoven (New York: Oxford University Press, 1998)) and James Hepokoski and Warren Darcy (Elements of Sonata Theory: Norms, Types, and Deformations in the Late-Eighteenth-Century Sonata (New York: Oxford University Press, 2003)) are either explicitly and avowedly ‘presentist’ or draw their evidence almost exclusively from the authors’ observations of historical repertories, offering little real engagement with historical theories.

Since partimenti include no explicit information on musical form and can mostly be described in terms of simple formal patterns, it is unsurprising that the partimento movement, focusing strongly on what Gjerdingen calls ‘moment-to-moment experiences’ (Music in the Galant Style, 424), has produced little discussion of musical form. Yet as soon as we extend historically inspired analysis based on thoroughbass and schemata to more complex examples of late eighteenth-century instrumental music, we do need categories to discuss musical form. Nothing speaks against the now fashionable combination of Gjerdingen's schemata with Caplin's ‘formal functions’ and the Elements of Sonata Theory by Hepokoski and Darcy. But it is somewhat disappointing to see the venture of a historically inspired Formenlehre set aside so quickly and so easily before its full potential has been tapped. And this, ultimately, is true for the project of historically inspired music analysis in general, which brings me to some concluding thoughts.

I remember precisely what crossed my mind while sitting at a piano and playing through Fenaroli's Regole for the first time a decade ago. I immediately felt that this was the definitive verdict on all those late nineteenth- and twentieth-century theories of harmony and counterpoint I had been brought up on: not because they were anachronistic (we knew that before) – no, because I realized that they are simply not sophisticated and precise enough to describe how this music works. In other words, the ultimate goal of a project of historically inspired analysis would not be to restrict ourselves to a set of historically ‘allowed’ terms, but to provide ‘a stimulus to consider a wider range of possibilities’ (Clive Brown, Classical and Romantic Performing Practice, 1750–1900 (Oxford: Oxford University Press, 1999), 2). The goal would also be a critical perspective capable of debunking modern ‘presentist’ (yet ageing) approaches that risk producing all the analytical problems they pretend to solve (to paraphrase Karl Kraus's dictum on psychoanalysis).

Only slowly did Kofi Agawu's question of ‘how we got out of analysis and how to get back in’ (Music Analysis 23/2–3 (2004), 267–286) return to my mind. It occurred to me that for a musicology that increasingly deconstructs, debunks and discards ‘old’, overly ‘textual’ and ‘immanent’ practices of ‘work-concept’-based analyses – but is not offering positive alternatives – the study of improvisational practices and related theoretical concepts may provide ways to get back into analysis. These would allow us to bring together research, analysis, improvisation and performance, to gauge anew the borderland between historical musicology and music theory – an opposition I saw collapsing before my eyes in 2007. Not all of this vision has been realized. We still have to rise to some of the manifold challenges that eighteenth-century music theory and a historically inspired practice of music analysis continue to pose. Partimenti and thoroughbass are not over. We have only just begun.