FORMAL DYNAMICS OF THE EIGHTEENTH-CENTURY TYPE 2 SONATA by CHELSEA NICOLE WRIGHT A DISSERTATION Presented to the School of Music and Dance and the Division of Graduate Studies of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy December 2021 DISSERTATION APPROVAL PAGE Student: Chelsea Nicole Wright Title: Formal Dynamics of the Eighteenth-Century Type 2 Sonata This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the School of Music and Dance by: Stephen Rodgers Chairperson Jack Boss Core Member Drew Nobile Core Member Jeffrey Librett Institutional Representative and Krista Chronister Vice Provost for Graduate Studies Original approval signatures are on file with the University of Oregon Division of Graduate Studies. Degree awarded December 2021 ii © 2021 Chelsea Nicole Wright iii DISSERTATION ABSTRACT Chelsea Nicole Wright Doctor of Philosophy School of Music and Dance December 2021 Title: Formal Dynamics of the Eighteenth-Century Type 2 Sonata Sonata form is arguably the most important form to develop in eighteenth-century instrumental music. In their 2006 treatise, Elements of Sonata Theory, James Hepokoski and Warren Darcy identify five sonata types prominent in the eighteenth century that interact with the “textbook” form taught in undergraduate classrooms today—what they call the “Type 3” sonata. Type 2 sonatas, or “sonatas without recapitulation,” are the subject of this dissertation. In these pieces, the return of the primary key near the end of the piece coincides with secondary theme material—seemingly passing over primary theme material. While the Type 2 form was extremely common in the mid-eighteenth century, its behaviors remain largely unexplored in the music-theoretical and analytical literature. The purpose of this project is to explore how features early in an eighteenth-century Type 2 movement interact with the moment of the primary key’s return later in the movement. I consider the role of main themes, cadences, and the layout of a piece’s development. Musical examples come from composers whose use of the Type 2 form is notable, from better-known composers such as Mozart and J.C Bach to lesser-known composers such as Johann iv Stamitz and Marianna D’Auenbrugg. By exploring how these features impact our hearing of a work’s tonal resolution, we pave the way for a deeper understanding about what makes these moments expressive and meaningful. v CURRICULUM VITAE NAME OF AUTHOR: Chelsea Nicole Wright GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene University of Colorado-Boulder, Boulder Sarah Lawrence College, Bronxville DEGREES AWARDED: Doctor of Philosophy, Music Theory, University of Oregon Master of Music, Music Theory, 2015, University of Colorado-Boulder Bachelor of Arts, Liberal Arts, 2011, Sarah Lawrence College AREAS OF SPECIAL INTEREST: Sonata Form Eighteenth-Century Keyboard Music PROFESSIONAL EXPERIENCE: Graduate Employee, University of Oregon, 2015–2020 Teaching Assistant, University of Colorado-Boulder, 2012–2014 vi ACKNOWLEDGMENTS Many thanks to the friends and colleagues who attended my research presentations, provided helpful feedback and encouragement, and who kept me company in the sometimes lonely act of writing this dissertation—especially in the midst of a global pandemic. I also wish to thank Professor Stephen Rodgers for his tremendous support, guidance, and patience as committee chair, advisor, and mentor. Thanks to my teachers throughout all my school years for your encouragement, inspiration, and wisdom. Finally, eternal thanks to my family for sticking with me every step of the way. vii For my grandparents. I carry you in my heart always. Para mis Abuelos. Siempre los llevo en mi corazón. viii TABLE OF CONTENTS Chapter Page I. INTRODUCTION: THE TYPE 2 SONATA ................................................................... 1 Analytical Goals and Questions ........................................................................................ 12 Historical Perspectives on the Type 2 Sonata .................................................................... 19 Modern Scholarship on the Type 2 Sonata ....................................................................... 23 Outline of this Dissertation ................................................................................................ 28 II. THE MEDIAL CAESURA ................................................................................................ 30 Medial Caesura Norms ..................................................................................................... 36 Exposition Types ............................................................................................................... 42 Two-Part Expositions ........................................................................................... 43 Continuous Expositions ........................................................................................ 43 Double Medial Caesura Effect ............................................................................. 50 Analysis: J.C. Bach, Keyboard Sonata in D Major, W A2/ii, Two-Part Exposition ........ 60 Analysis: J.C. Bach, Keyboard Sonata in B-Flat Major, W A1/i, Two-Part Exposition ......................................................................................................... 64 Analysis: J.C. Bach, Keyboard Sonata in C minor, W A6/i, Continuous Exposition ...... 68 Analysis: J.C. Bach, Keyboard Sonata in C minor, W A8a/iii ......................................... 78 III. THE DEVELOPMENT ................................................................................................... 91 ix Chapter Page Characteristics ................................................................................................................... 98 Thematic ............................................................................................................. 103 Rhetorical .......................................................................................................... 105 Tonal ................................................................................................................. 112 Rotation and the Type 2 Sonata ....................................................................................... 114 Johann Stamitz and the Type 2 Sonata ............................................................................ 121 Analysis: Johann Stamitz, Symphony in G major, G-2/iv ................................................ 122 Analysis: Johann Stamitz, Symphony in E-Flat Major, Eb-1/i ......................................... 125 IV. THE TONAL RESOLUTION ........................................................................................ 131 A Positive Approach to Tonal Resolutions ....................................................................... 137 The “Reversed Recapitulation” ........................................................................................ 141 Process and the Tonal Resolution ..................................................................................... 146 Analysis: Marianna D’Auenbrugg, Keyboard Sonata in E-flat Major, i .......................... 152 Analysis: Wolfgang Amadeus Mozart, Symphony in D Major, K. 504/i (“Prague”) ....... 159 V. CONCLUSION ................................................................................................................. 176 REFERENCES CITED .......................................................................................................... 180 x LIST OF EXAMPLES Example Page 1.1. Boccherini, Cello Sonata G.1/i, Exposition ......................................................... 1 1.2. Boccherini, Cello Sonata G. 1/i, Development ................................................... 9 2.1. J.C. Bach, Keyboard Sonata W A2/ii, mm. 8–9. ................................................ 30 2.2 J.C. Bach, Keyboard Sonata W A2/i, mm. 17–19 ............................................... 61 2.3 J.C. Bach, Keyboard Sonata W A2/i, mm. 33–36 ............................................... 63 2.4 J.C. Bach, Keyboard Sonata W A1/i, mm. 10–14 ............................................... 65 2.5 J.C. Bach, Keyboard Sonata W A1/i, mm. 54–57 ............................................... 66 2.6 J.C. Bach, Keyboard Sonata W A6/i, mm. 1–19 ................................................. 70 2.7 J.C. Bach, Keyboard Sonata W A6/i, mm. 40–56 ............................................... 76 2.8 J.C. Bach, Keyboard Sonata W A8a/iii, mm. 1–13 ............................................. 79 2.9 J.C. Bach, Keyboard Sonata W A8a/iii, mm. 14–23 ........................................... 81 2.10 J.C. Bach, Keyboard Sonata W A8a/iii, mm. 22–29 ......................................... 83 3.1. Stamitz, symphony in D major, D-2/i ................................................................. 92 3.2. Stamitz, Symphony in G major, G-2/iv, mm. 1–41 ............................................ 123 4.1. D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 1–19 ............................ 152 xi Example Page 4.2. D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 14–36 .......................... 153 4.3. D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 30–56 .......................... 154 4.4. D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 57–76 .......................... 155 4.5. D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 57–76 .......................... 157 4.6. Mozart, Symphony in D Major, K. 504 “Prague” (i), mm. 37–43 ...................... 165 4.7. Mozart, Symphony in D Major, K. 504 “Prague” (i), mm. 91–100 .................... 166 4.8. Mozart, Symphony in D Major, K. 504 “Prague” (i), mm. 140–162 .................. 168 xii LIST OF FIGURES Figure Page 1.1. Formal Diagram of the Type 3 Sonata ................................................................ 3 1.2. Comparison of rounded and balanced binary with the Type 2 Sonata ............... 5 1.3. Spectrum between binary form and sonata form ................................................. 24 2.1. MC types in J.C. Bach’s Type 2 keyboard sonata movements ............................ 40 2.2. Formal layout of a two-part exposition ................................................................ 43 2.3. Formal layout of continuous expositions subtypes 1 and 2 .................................. 46 2.4. Formal layout of J.C. Bach W A6/i ..................................................................... 71 2.5. Formal plan of J.C. Bach W A6/i, exposition ...................................................... 73 2.6 J.C. Bach W A8a/iii, readings of Rotation 1 ........................................................ 85 2.7 J.C. Bach W A8a/iii, readings of Rotation 2 ........................................................ 86 3.1. Stamitz, Symphony G-2/iv: vi: PAC ends the development ................................ 124 3.2. Stamitz, Symphony Eb-3/i Exposition Themes .................................................. 126 4.1.Type 3 sonata with reversed recapitulation ........................................................... 142 4.2 Type 2 sonata with P-based coda .......................................................................... 142 4.3 Formal layout of Mozart, Symphony K. 504/i ..................................................... 163 xiii CHAPTER I INTRODUCTION: THE TYPE 2 SONATA Let’s begin with some music. Example 1.1 shows the exposition of the first movement of Luigi Boccherini’s Cello Sonata in F major, G. 1. Example 1.1: Luigi Boccherini, Cello Sonata in F major, G. 1/i, exposition 1 On a large-scale level, this exposition does a lot of the things a Classical exposition is “supposed” to do: we have thematic material in the tonic, thematic material in the dominant, and the whole thing ends decisively before the development begins. These are characteristics that could belong to sonata forms or binary forms, early eighteenth- or early-nineteenth century compositions. Upon closer analysis, the exposition gets more complicated. Where is the second theme? Does it begin in the second half of m. 7? Or perhaps mm. 7–11 are transitional and the second theme begins in the minor dominant, in m. 11. But then again, could S begin all the way in m. 17? If one expects a “textbook” sonata form to unfold, the rest of the movement may also seem strange or difficult to parse. Figure 1 shows the layout of such a “textbook” form, as Hepokoski and Darcy define it in their 2006 monograph Elements of Sonata Theory: Norms, Types, and Deformations in the Late-Eighteenth-Century Sonata—what we can reasonably 2 expect given both the prevalence of this form in the repertoire and its overwhelming dominance in historical, theoretical, and analytical discussions of sonata form. Hepokoski and Darcy call this textbook sonata form the “Type 3 sonata”; Figure 1.1 reproduces their Figure 2.1b. Figure 1.1: Formal Diagram of the Type 3 sonata, from Hepokoski and Darcy’s Elements of Sonata Theory1 Upon listening, it quickly becomes apparent that Boccherini’s G. 1 does not reflect the structure of this textbook sonata form. For one thing, there is a section that is clearly development, and a section that is clearly a return of thematic return in the tonic (and, for the most part, in its expositional form), but there is not a clear seam between the two. The F minor passage in m. 38 occurs almost exactly as it does in the exposition (with the 1 James Hepokoski and Warren Darcy, Elements of Sonata Theory: Norms, Types, and Deformations in the Late- Eighteenth-Century Sonata (New York: Oxford University Press, 2006), 17. 3 exception of m. 43, which is an expansion and rhythmic alteration of the motive in m. 16). And yet the cadence that precedes it in m. 38 (if one even chooses to identify it as such) is extremely weak. Can this really be the line between development and recapitulation? This question might not be so pressing but for the fact that, aside from whether the theme is tonally and thematically stable, it comes from well within the exposition; in other words, when the tonic is restored, it is not accompanied by the primary theme. In fact, the primary theme does not return at all for the rest of the movement. This movement could be defined in a variety of ways, with names such as “binary sonata form,” and “polythematic binary form.” A binary sonata form, as defined by R.M. Longyear, is characterized by the following features: the second half of the movement starts with primary theme material in a key other than the tonic, and the return of the tonic is accompanied by the return of secondary theme material.2 Similarly, a polythematic binary form, as defined by Eugene K. Wolf, “introduces thematic differentiation into simple binary form….It often accentuates in various ways a coordinated return of both S and the tonic key.”3 It is the “coordinated return” of the tonic and secondary theme material characteristic of these formal descriptors that renders them appropriate for G. 1. The Type 2 sonata (or “sonata without recapitulation”), as defined by Hepokoski and Darcy in Elements of Sonata Theory, also describes the structure of G. 1. The Type 2 sonata’s most salient feature is the tonic’s return post-development with 2 R.M. Longyear, “Binary Variants in Early Classic Sonata Form,” Journal of Music Theory 13, no. 2 (1969): 164-165. 3 Eugene K. Wolf, The Symphonies of Johann Stamitz: A Study in the Formation of the Classical Style (Utrecht: Bohn, Scheltema, & Holkema, 1981), 139. 4 secondary rather than primary theme material.4 Figure 1.2 shows diagrams of a binary form and Hepokoski and Darcy’s diagram of the Type 2 sonata. Figure 1.2: Comparison of rounded and balanced binary form with the Type 2 sonata a: Rounded Binary form ||: A :||: BA' :|| b: Balanced Binary form5 ||: A1 + B1 :||: C + B1 :|| c: Formal Diagram of the Type 2 sonata, from Hepokoski and Darcy’s Elements of Sonata Theory6 Each term has slightly different connotations from the rest, and not all are equally applicable to this movement. Rounded binary, which ends with an altered version of what we would consider in a sonata form to be primary theme material, is, right off the 4 Hepokoski and Darcy, Elements of Sonata Theory, 344. 5 Hepokoski and Darcy, Elements of Sonata Theory, 355, after Douglass M. Green, Form in Tonal Music: An Introduction to Analysis (New York: Holt, Rinehart and Winston, Inc., 1965) 76, 91. 6 Hepokoski and Darcy, Elements of Sonata Theory, 354. 5 bat, not a good fit for G. 1; in G. 1 it is material from later in the exposition that returns in the tonic, including secondary theme material; rounded binary looks and sounds more like the “textbook” Type 3 sonata as Hepokoski and Darcy define it. Balanced binary form and the Type 2 sonata are better fits—balanced binary features an “end-rhyme” in which material from the end of the first section returns at the end of the second section. Even so, Douglass M. Green’s balanced binary layout features a second section that opens with new material, while G. 1’s second section opens with primary theme material in the dominant.7 A Type 3 reading, which would require accounting for the “omission” of the recapitulation of primary theme material in the tonic, is an even less convincing option, while a Type 2 reading becomes increasingly hard to ignore. This is before one even considers historical factors, such as the fact that the kind of alterations— “deformations,” in Hepokoski and Darcy’s terminology—required to read G. 1 in Type 3 terms are very rare for pieces in Boccherini’s time. After all is said and done, the most straightforward analytical framework for this movement is to read or hear it as a binary sonata form. Distinctions between “simpler” binary forms and sonata-like structures can be both difficult to determine as well as not necessarily of great analytical significance. Historically, scholars have taken different approaches to mapping binary (and ternary) forms onto sonata forms.4 The line between binary form and sonata form is porous, especially in galant sonatas like Boccherini’s. Binary sonata form and polythematic binary form are often used to refer to the form of the first movement of an eighteenth-century 7 Douglass M. Green, Form in Tonal Music: An Introduction to Analysis (New York: Holt, Rinehart and Winston, inc., 1965), 76. 6 multi-movement work, especially a solo sonata or symphony, that does not have a complete recapitulation of the exposition’s thematic material, as in G. 1; it is a compromise between the clear binary structure and its inter- and intra-movement dialogue with sonata form. These terms, however, and the perceived lack of recapitulation they reflect, also often come with a “demotion” (if unintentionally) from the typical sonata-form-first-movement to identify a movement as “merely” a binary form, though with some gestures toward its sonata-form-like qualities. After all, doesn’t a sonata form require an exposition, development, and recapitulation? The Type 2 sonata is the most recent of the formal descriptions above: one which captures the formal, generic, and historical relationship between these pieces and sonata form while also distinguishing it from the more familiar, “textbook” Type 3 sonata (which has an exposition, development, and recapitulation as taught in most undergraduate classrooms). It is also associated with a single theoretical framework—Hepokoski and Darcy’s Sonata Theory. According to their theory, sonatas can be in dialogue with five sonata types (the “foundational axioms” that make a sonata a sonata will be explored later in this chapter): the Type 1 sonata, often described as a “sonata without development”; the Type 3 sonata, the “textbook” sonata with exposition, development, and recapitulation; the Type 4 sonata, the sonata-rondo; and the Type 5 sonata, the concerto. Hepokoski and Darcy would define G. 1 above as a Type 2 sonata, a “sonata without recapitulation.” In these sonata forms, the development is not followed by a recapitulation of all the exposition’s material in the tonic key; rather, only the second theme group returns in tonic. Sonata Theory offers a framework by which these forms 7 can be both heard in dialogue with sonata form and also differentiated from the “textbook” sonata form with exposition, development, and recapitulation. There are many other expositions like this in Boccherini’s oeuvre, and similar questions arise when analyzing other movements from the early and mid-eighteenth century (albeit generally manifested in less extreme ways). In other words, while such expositions may be challenging to parse given general theoretical and analytical bias toward late-eighteenth- century Viennese compositions, they are not really aberrant for their own time.8 In these movements, formal clarity doesn’t just come from the moment of tonal resolution (what others might think of as the S-based “recapitulation”) but it is actually a result of the formal organization of the rest of the movement. G. 1 shows the Type 2 in all its glory insofar as it shows how illogical it is to try to divide a piece like this into development and recapitulation. While it is an extreme example (compared to the Type 2s of other composers, though not so unusual among Boccherini’s), it is also a fair representative of the fundamental analytical importance of distinguishing Type 2 and the “textbook” Type 3 sonata from one another. Because of the complexity of its expositional structure, it is a good demonstration of the analytical questions that might inspire the analyst to see the Type 2 sonata as a form defined by a host of formal processes rather than defined as a single moment in time (i.e., the moment of so-called “tonal resolution”). In order to make sense of the various formal functions at play in Boccherini’s G. 1, we need a cipher of sorts. This sonata provides one in the form of the movement’s 8 Rebecca Long’s dissertation explores form in Boccherini’s compositions (through the context of extrinsic phrases); she also analyzes G. 1 as a Type 2 sonata. See Rebecca J. Long, “Extrinsic Phrases in Early- Classical Sonata Forms,” Ph.D. dissertation, University of Massachusetts Amherst, 2018. For Long’s analysis of G. 1, see pp. 97-112. 8 second half. In other words, understanding the formal organization of the second half of the movement sheds light on the formal organization of the exposition. Example 1.2 shows the development of this sonata movement. Example 1.2: Luigi Boccherini, Cello Sonata in F major, G. 1/i, development 9 The development begins with the first theme in the dominant, a very typical opening for developmental space in galant sonata forms. Measure 29 launches a varied iteration of the theme from m. 7 that transforms from the development’s C major opening to the dominant of the global tonic F major. This passage ultimately spins out beyond the scope of its expositional version, but the rhythmic and gestural content resonates clearly with its m. 7 counterpart. That it is set over the dominant (including the chordal seventh) amplifies the transitional quality of this theme, something we can retrospectively interpret in the exposition. The dominant, while prolonged from mm. 30-33, does not give way to a permanent, secure return to F major. Instead, after a momentary respite for a half cadence in m. 33, this theme resumes and ultimately spans nearly eight measures, double the length of its expositional version. From this moment on, the sonata closely follows the exposition, this time in the tonic—though m. 43 expand and rhythmically alters the motive of m. 16. All of this information can help us sort out the exposition’s thematic and rhetorical trajectories. Let’s begin with what is clear: mm. 1–7 form a single musical thought; there 10 is a sense of arrival with the half cadence in m. 7, and the pick-ups into m. 8 begin a new thought. The passage beginning in m. 8 has a slow harmonic rhythm of one chord change per measure and increased surface rhythm with a constant stream of sixteenth- note triplets. This thought ends on the downbeat of m. 11 and a new thought begins with the pick-up to measure 12. The next segment arrives following a V: HC in m. 17. and is the final musical thought of the exposition. Each of these themes is clearly differentiated from one another by means of phrase endings, texture, and rhythmic intensity. What is not entirely clear is how each of these themes functions within the context of the exposition. At issue are the themes of m. 8 and m. 11, which have tonal and melodic characteristics of transition sections: each exists in a tonal area clearly centered around C (C major in m. 8 and C minor in m. 11), but the pedal-point on ^5 of C major in the bass maintains the expectation of a tonal arrival. The melodic units begun in mm. 8 and 11 are also not entirely stable. The octave leaps on ^5 of C major in m. 8 create a yo-yo effect, and the sudden increase in surface rhythm also undermines the sense of stability. In m. 11, the sudden change in mode and surface rhythm is also destabilizing. What is particularly interesting in this case is that all of this transitional material happens after the half cadence that normally signals the arrival of the second theme (what Hepokoski and Darcy call the medial caesura, “the brief, rhetorically reinforced break or gap that serves to divide an exposition into two parts [P and TR; S and C].”9 In the second half of the movement, when the m. 8 theme returns in 9 James Hepokoski and Warren Darcy, Elements of Sonata Theory: Norms, Types, and Deformations in the Late- Eighteenth-Century Sonata, (New York: Oxford University Press, 2006), 24. See also Hepokoski and Darcy, “The Medial Caesura and Its Role in the Eighteenth-Century Sonata Exposition,” Music Theory Spectrum 19 (1997): 115–54. 11 m. 29, its transitional qualities are amplified by increased motion in the bass (including chromatic tones), as well as chromatic tones in the melody. It is fairly clear that this passage is still in development-space, but less so when the minor theme of m. 11 returns in m. 38. Is this still part of the development, albeit as a passage that clearly presages the tonal resolution to come? Or does it indeed commence the tonal resolution, the modal resolution to occur later in m. 44? That the development and tonal resolution together form a single span that cannot be readily divided confirms the movement’s Type 2 form. It is not a single moment that makes this a Type 2 rather than a Type 3 sonata; we can relate the continuity of the movement’s second half to the blurred boundaries of the exposition. It is this link—between the exposition and the movement’s development-tonal resolution seam—that clarifies how the music hangs together in a coherent structure. Analytical Goals and Questions To date, much of the theoretical focus on the Type 2 sonata has focused on the identity of these pieces: what are they, really? Binary forms, sonata forms, or something else? The issue of which label to choose is important insofar as it 1) places a work in dialogue with other, similar works, and 2) reflects something about the piece’s generic references and stylistic or expressive nature. What has received less scholarly attention is the behavior of these pieces, or how their constituent parts unfold and interact with one another over the course of the entire form. This is especially true for Formenlehre regarding repertoire outside the realm of late- eighteenth-century Viennese composers (read: Haydn, Mozart, and Beethoven). When 12 other repertoire, including music of the galant, is analyzed, it is often through lens of late- eighteenth-century norms, regardless of whether the music itself lends itself to analysis in that way. While some of the general principles derived from the sonata forms of Haydn, Mozart, and Beethoven can be applied effectively to galant music, sometimes some modification is needed. A clear example of this, which I will discuss further in Chapter II, concerns medial caesura norms. While half cadences in the dominant key predominate in late Classical sonata forms, half cadences in the tonic key are more common in galant repertoire. The cadence serves the same function regardless—dividing the exposition into two parts—though the norms of how this cadence unfolds change over time. The purpose of this dissertation is to explore how elements that are not unique to the Type 2 sonata interact with this formal type in particularly expressive ways. How do the implications of an expositional event (for instance, the end of P or the start of S) contribute to a movement’s “Type-2-ness,” as defined by its bi-rotational structure with a continuous trajectory between the start of the development and the end of S or the closing section? How might the effect of the medial caesura differ when it occurs between development and tonal resolution as opposed to when it occurs securely within the bounds of a recapitulation? This dissertation explores the formal processes that arise from the interaction of such events with the large-scale development-tonal resolution of a Type 2 movement’s second half. Each chapter focuses on a different component of sonata form and what role that component might play in the formal and expressive narrative of a Type 2 sonata specifically. Moreover, I follow Sonata Theory and consider the Type 2 sonata a norm in and of itself rather than a deformation of something else (the Type 3 sonata). While, even in the galant, the Type 3 sonata appeared with greater frequency, 13 there is no indication that the Type 2 sonata was deemed a lesser form of the Type 3 sonata. In this dissertation I focus more on how formal events interact with one another than on assigning formal labels to a given work. Rather than emphasize a binary between Type 2 and Type 3 sonatas, I instead focus on exploring how the movement expresses characteristics of each—or in some cases, both—formal types. Similarly, a binary form (whether judged as such based on historical, generic, or formal grounds) that exhibits some Type 2 qualities may merit further consideration in this project. Such a flexible approach better reflects the flexible and rapidly changing norms of the galant. Like many sonatas beyond the scope of the Classical style, Type 2 sonatas demonstrate how porous and fragile the divide is between “sonata” and “not-sonata.” By taking galant repertoire as a starting point, the dissertation asks these questions in the context of a musical style in which the Type 2 form is found with great frequency. It is not unusual or aberrant in this repertoire, to a much greater degree than in the late- eighteenth-century repertoire upon which Sonata Theory is based. I claim that the Type 2 sonata should be understood not as an under-formed, lesser version of the Type 3 sonata but as its own distinct sonata type.10 Even barring any value judgment, the processes that drive a Type 2 sonata should be explored on their own terms, rather than only in relation to their Type 3 counterparts. In that spirit, I argue that Type 2 is not defined by one moment but by how that moment interacts with the rest of the musical 10 As I shall explore in Chapter IV and Chapter V, however, boundaries between formal types remain somewhat porous, leading to formal hybrids and type-conversions. 14 events and trajectories of the movement—and that these trajectories are necessarily different from their Type 3 counterparts. In Sonata Theory, the concept of sonata type is linked closely to the idea of rotation. One of the central tenets of Sonata Theory is that sonata form is constructed from a series of rotations through a referential thematic layout established in the exposition. The Type 3 sonata is a tri-rotational form: the exposition, development, and recapitulation each constitute a rotation through an ordered sequence of themes. The idea of rotation clearly demonstrates the parallel between the exposition and the recapitulation that other theories of sonata form have long described. That the development of a sonata form is also essentially rotational is less obvious to the eye and ear. I explore this perceived problem in Chapter III. For now, suffice it to say that, especially in galant sonatas, development sections are more overtly and consistently rotational than their late- eighteenth-century counterparts. My approach draws primarily from Hepokoski and Darcy’s Elements of Sonata Theory and Hepokoski’s A Sonata Theory Handbook.11 While Hepokoski and Darcy’s theory is derived from late-eighteenth-century compositions, many of its principles are evident in galant compositions from the preceding centuries. Sonata Theory highlights the rotational and parallel two-part elements of these compositions and is well-suited to reveal how sonata-processes associated with the Type 3 sonata are in play as well. While Sonata Theory’s concept of rotation as an essential part of sonata form has been contested, especially its applicability to nineteenth-century works (notably by Paul Wingfield12) in 11 James Hepokoski, A Sonata Theory Handbook (New York: Oxford University Press, 2020). 15 galant sonata forms, Sonata Theory’s principles of rotation are quite evident. This is especially true for how the Type 2 sonata’s birotational form is laid out—that the second part of the sonata (development and tonal resolution) constitutes a single trajectory through the exposition’s referential thematic layout is very salient. There is an overwhelming preference shown among these works to begin the development with P in V (or, in the relatively few minor-mode sonata forms of the period, in III), and even when there are internal variations on the referential material (or altogether new material), the overall movement from P to S generally remains clear. The principle of rotation also allows the analyst to distinguish between sonata types and between different types of return. Rotation is fundamentally about thematic return and about situating thematic return within a larger rhetorical trajectory; where these returns occur is as important as the fact that they return at all. This provides an analytical framework for understanding how the tonic return of S in a Type 3 sonata is fundamentally different from that of a tonic return of S in a Type 2 sonata. While the question of whether the Type 2 sonata (and Sonata Theory more generally) is an appropriate lens for nineteenth-century works remains for another day, that it is an appropriate lens for galant sonata forms is harder to dispute. In contrast to the Type 3 sonata’s three rotations, Type 1 and Type 2 sonatas are bi-rotational. In a Type 1 sonata, the first rotation constitutes the exposition and the second the recapitulation. In a Type 2 sonata, the first rotation constitutes the exposition and the second both the development and what Hepokoski and Darcy call the “tonal 12 Paul Wingfield, “ ‘Beyond Norms and Deformations’: Towards a Theory of Sonata Form as Reception History,” Music Analysis 27, no. 1 (2008): 148-153. 16 resolution,” the moment at which the tonic is secured following the development (note that the musical examples in this dissertation are mostly in major keys, and development sections are relatively brief compared to those of the late-eighteenth-century sonata).5 In a Type 3 sonata, the tonal resolution coincides with the start of the recapitulation where P is stated in the tonic. In a Type 2 sonata, the tonal resolution coincides with S in the tonic. From there, the rotation continues according to the referential layout; the tonic is maintained, and P never returns within sonata space.6 As Hepokoski and Darcy note, there is no essential distinction between the exposition of a Type 2 sonata and that of a Type 3 sonata. In other words, in theory a given exposition could belong to either a Type 2 or a Type 3 sonata; it is not possible to tell which will unfold based on the exposition alone.13 The start of the development does not necessarily reveal the formal type either; I will discuss exceptions in Chapter III. If the development constitutes a complete rotation, that is one signal that a Type 3 sonata is in play; once that rotation is complete, a new rotation must be launched—in this case, the recapitulation. In a Type 2 sonata, the second half of the movement constitutes a single rotation that includes the development and the tonal resolution. In other words, P (and TR, if present) is the basis for the development, and the rest of the referential layout, beginning with S, sounds in the tonic. Given the prevalence of the Type 3 sonata amongst the other sonata types—not just in the eighteenth century, but in the nineteenth as well—it is difficult not to immediately compare the Type 2 sonata to the Type 3 model, to think of it as a variation on the Type 3 layout. This bias is exacerbated by the focus in music studies on late- 13 Hepokoski and Darcy, Elements of Sonata Theory, 369. 17 eighteenth-century repertoire, where the Type 3 sonata’s prevalence far outstrips that of the Type 2 sonata. The problem arises when considering galant repertoire with this same lens. While the Type 3 sonata still predominates among sonata types in the galant period, the Type 2 sonata occupies a much larger majority among galant sonata forms. Nearly half of J.C. Bach’s keyboard sonata movements in sonata form are Type 2 sonatas, and Type 2 sonatas make up the overwhelming majority of Johann Stamitz’s first movements.14 The Type 2 sonata is an excellent place to begin shifting our approach to galant repertoire, letting it come out from the shadow of the more dominant Type 3 sonata. Despite its rarity in late-eighteenth-century repertoire (and its generally smaller scale), the Type 2 sonata is neither a lesser nor a less common form in galant repertoire. Indeed, to ignore it is to ignore part of what distinguishes the galant from the late Classical style. In general, the principles of Sonata Theory apply well to mid-century repertoire. In other words, the trajectories and patterns identified in Sonata Theory are evident in this earlier repertoire. Differences arise more in terms of specific norms and analytical values—for instance, the shift in relative importance of the recapitulation in mid- versus late-eighteenth-century sonatas, to be discussed further in Chapter IV. For now, suffice it to say that, despite its derivation from later material, the basics of Sonata Theory reflect well the mechanics of music from the preceding few decades given the modification of certain norms and expectations. 14 Eugene K. Wolf, The Symphonies of Johann Stamitz: A Study in the Formation of the Classical Style, (Utrecht: Bohn, Scheltema & Holkema, 1981), 139. 18 Historical Perspectives on the Type 2 Sonata Elements of Sonata Theory includes the most extensive and systematic exploration of the form Hepokoski and Darcy came to call the Type 2 sonata. There is little reference in late- eighteenth- and early-nineteenth-century writings on musical form to this kind of structure. Francesco Galeazzi, in his Elementi teorico-practici de musica (1791) writes that: The Reprise succeeds the Modulation. However remote the Modulation is from the main key of the composition, it must draw closer little by little, until the Reprise, that is, the first Motive of Part 1 in the proper natural key in which it was originally written, falls in quite naturally and regularly. If the piece is a long one, the true Motive in the principal key is taken up again, as it has been said, but if one does not want to make the composition too long, then it shall be enough to repeat instead the Characteristic Passage transposed to the same fundamental key....If the second method has been used—that is, the reprise of the Characteristic Passage—then the Modulation shall be ended on the dominant of the key, in order to start then the Characteristic Passage in the main key; and also in this case it is good practice to touch upon somewhere, though slightly, the modulation to the subdominant of the key.15 Heinrich Christoph Koch does not single out the “Type 2” sonata in his Versuch einer Anleitung zur Composition (1782–1793), but his description of formal options for the symphony does include the possibility of the Type 2 sonata. He introduces this possibility in the context of the symphony, later noting that “the sonata assumes all the forms which already have been described before in connection with the symphony.” Koch writes, “the last period of our first allegro, which is devoted above all to the main key, most frequently again begins with the theme in [the main] key, but occasionally may also start with another main melodic idea” (italics mine).16 One important difference between Koch’s formulation and 15 Bathia Churgin, “Francesco Galeazzi’s Description (1796) of Sonata Form,” Journal of the American Musicological Society 21 (1968), 195–196. 16 Heinrich Christoph Koch, Versuch einer Anleitung zur Composition, trans. Nancy Kovaleff Baker (New Haven: Yale University Press, 1983). For the first quote, see §103, on p. 201. For the second, see §110 on p. 204. 19 Hepokoski and Darcy’s is that he does not require post-development tonal return to coincide with S material; it can be anything aside from P. Just as sonata forms themselves are products of their times, so are descriptions of formal structures shaped by the prevailing musical values and expectations of their writers. Moreover, the role of genre in the late-eighteenth-century descriptions from which we might derive a historical approach to form should not be ignored. In other words, consider that what we learn from Koch about how he conceived of what we call sonata form is laid out in terms of “the nature and arrangement of the most common compositions,” which include the symphony, sonata, duet, and trio.17 Other treatises that address sonata form as a large- scale form (such as those of A.B. Marx and Antoine-Joseph Reicha) are filtered through the norms of late-eighteenth-century style.18 In the context of Type 2 sonatas specifically, there are a few shifts in ideas about sonata form that are particularly important: the shift in emphasis on the drama of the onset of the development to the onset of the recapitulation (which is interconnected with a shift in emphasis from the binary to ternary aspects of the form). We should take care in assuming too much cause and effect in this area, and instead focus on the Type 2 sonata’s shifting prevalence as these sonata-form perspectives are changing in the late eighteenth and early nineteenth centuries. While my approach to the Type 2 sonata in this dissertation does not ignore historical considerations—indeed, the stylistic conventions of the galant, even and perhaps especially when they differ from those of the late eighteenth 17 Heinrich Christoph Koch, Versuch, 165-213. 18 A.B. Marx, Die Lehre von von der musikalischen Komposition, practisch-theoretisch, vol. 1, trans. August H. Wehrhan (London: R. Cocks, 1852); Antoine-Joseph Reicha, Traité de haute composition musicale (Paris: Zetter, 1824-1826). 20 century, are central to my approach to the Type 2 sonata—it is also grounded primarily in modern Formenlehre. As twenty-first-century scholars, we have the opportunity to take a bird’s-eye view of eighteenth-century music unavailable to contemporaries (just as a bird’s-eye view of our own time will belong to our successors). While formal descriptions cannot be purely objective or detached from our own biases, we can still consider relationships and similarities between works without implying cause and effect. The heyday of the Type 2 sonata came in what Mary Sue Morrow calls “a period in flux,” between the late Baroque and high Classical periods.19 The norms and expectations of late-eighteenth-century sonata form have not yet come into full effect in the 1750s and 1760s. This is the time of the galant, of musical styles with origins in Italian theater spanning the decades before and after 1750. Definitions of galant style abound, but both C.P.E. Bach and Heinrich Christoph Koch place it in contrast with the “learned” style.20 To lump galant music in with either Baroque or classical music in history or analysis does not do justice to these nuances. An important predecessor to the galant Type 2 sonata is the body of keyboard sonatas by Domenico Scarlatti.21 The vast majority of these single-movement sonatas fall under the category of what Douglass M. Green has called “balanced binary form,” described near the start of this chapter, and which Hepokoski and Darcy define as ||: A1 19 Mary Sue Morrow, “Sketch for a History of the Eighteenth-Century Symphonic Repertoire,” in The Eighteenth-Century Symphony, vol. 1 of The Symphonic Repertoire, ed. by Mary Sue Morrow and Bathia Churgin, (Bloomington: Indiana University Press, 2021), 780. 20 C.P.E. Bach, Essay on the True Art of Playing Keyboard Instruments, trans. William J. Mitchell (New York: W.W. Norton & Company, 1949) 16; Heinrich Christoph Koch, Musikalisches Lexikon (Frankfurt, 1802), trans. by Leonard G. Ratner, Classic Music: Expression, Form, and Style (New York: Schirmer Books, 1980) 23. 21 + B1 :||: C + B2 :|| (see figure 1.2b).22 Thematically, the balanced binary layout can almost be considered a generalized model of the Type 2 sonata—albeit with an important potential difference—one which does not account for the cadential or tonal parameters outlined in Sonata Theory. That the second section of a balanced binary begins with new, contrasting material belies the principle of rotation that underpins the Type 2 sonata. While it is possible to read such a contrasting section (marked C in this diagram, and not to be confused with “closing” material) as an example of what Hepokoski and Darcy call “writing over,” in which new material occupies space that is nonetheless heard as rotational (standing in for A, in other words, rather than reflecting a fundamentally different structure), the effect of such a substitution still undermines the listener’s perception of rotation.23 Moreover, writing over occurs substantially less in Type 2 sonatas than in its Type 3 counterparts—which can often be connected more directly to rounded binary forms. Regarding Scarlatti’s sonata forms in particular, W. Dean Sutcliffe remarks on the relative un-themeliness of Scarlatti’s first themes, whereas the second and closing themes of these sonatas are relatively cohesive: “Scarlatti makes ‘architectural’ capital out of...musical imagery, dissonance, syntactical style or keyboard sonority.”24 This is useful in the analysis of galant works whose expositional structures, when viewed (purely) thematically, are difficult to parse. In other words, even in cases of incipient medial caesuras (where there may be some separation that is defined by a weak cadence or other 22 Green, Form in Tonal Music, 76, 91. 23 Hepokoski and Darcy, Elements of Sonata Theory, 355-358. 24 W. Dean Sutcliffe, The Keyboard Sonatas of Domenico Scarlatti, (Cambridge: Cambridge University Press, 2008), 321-322. 22 contrast), are there elements that suggest the division of the exposition into two parts by more “stylistic” means? This is not to say the architectonic features of Type 2 as described in Elements are irrelevant, but rather that they form only a part of the picture of Type 2- ness (hence the wide variety of movements that all fall under the Type 2 umbrella). While possible generic origins and a precisely-defined relationship between the Type 2 and Type 3 sonatas are debatable, the Type 2 sonata, as Hepokoski and Darcy put it, “coexisted as a viable option standing side-by-side with [the Type 3 sonata], even though it was one less often adopted.”25 Whatever the origin of the forms, their shapes— bi-rotational versus tri-rotational—as well as their common use (especially in first movements) lend themselves to analysis through slightly different lenses that are nonetheless related. Modern Scholarship on the Type 2 Sonata Modern scholarship on the Type 2 sonata takes several forms, which we can first organize into pre- and post-Sonata Theory categories. As noted at the start of the chapter, theoretical conceptions of this form before Sonata Theory defined it using some different parameters, and the terms associated with the form reflect these differences. For instance, terms like those presented at the start of this dissertation, like “binary sonata form” and “polythematic binary form” seem to suggest that the apparent absence of a recapitulation of P material (assuming that such a recapitulation is an essential part of sonata form) distinguishes these forms from “true” sonata forms; considering a spectrum from binary 25 Hepokoski and Darcy, Elements of Sonata Theory, 366. 23 form to sonata form, these terms suggest that a Type 2 sonata lies closer to the binary side than that of sonata form (see Figure 1.3). Figure 1.3: Spectrum between binary form and sonata form Another approach to the Type 2 sonata considers it to be an altered version of a sonata form, whether using the Sonata Theory terminology of the Type 3 sonata or not. This view is considerably more common in the context of nineteenth-century movements than those of the eighteenth-century—and especially those of the galant—but not exclusively so. Consider, for instance, William Caplin’s description of “deletion of the main-theme opening” in the context of recapitulatory alterations: Some recapitulations delete the opening material of the main theme or even the entire theme. At times the transition may be eliminated as well, and the recapitulation begins directly with the subordinate-theme area. Although deviant in the high classical style, this procedure is normative in midcentury works and has its roots in baroque binary dance forms.26 Caplin goes on to question whether thinking of such returns as recapitulations is appropriate, ultimately determining that “since it is so traditional to label the main section following the development a recapitulation, the practice can still be maintained despite…theoretical concerns;” even accounting for this reserve, however, Caplin views 26 William E. Caplin, Classical Form: A Theory of Formal Functions for the Instrumental Music of Haydn, Mozart, and Beethoven (New York: Oxford University Press, 1998), 173. 24 this “deletion” as a modification of a more typical practice—that of a sonata with complete recapitulation.27 These contrasting perceptions of what Sonata Theory calls the Type 2 sonata come from historical and more theoretical considerations. In the midst of the emergence of the high Classical style (and in the decades preceding it), there is no single moment when the designation of the term “sonata form” becomes analytically appropriate; rather, sonata form is a formal “backdrop [of] complex sets (or constellations) of flexible action- options, devised to facilitate” dialogue between that backdrop and individual compositions”;28 while historical considerations inform that dialogue (a piece composed in 1760 will have a stronger connection to sonata form than one composed in 1710, a piece on the cusp of the galant that reflects a similar sequence of events and tonal-thematic relations may be profitably analyzed in the context of such a dialogue as well. Note that the above “alternative” perspectives on the Type 2 sonata—those of binary sonata form and altered “Type 3” sonata form—consider the form from different historical-theoretical perspectives. A nineteenth-century movement with what I would call Type 2 characteristics is more likely to be analyzed as an altered “Type 3” than as a binary form; the opposite is true for early galant movements. This is because of the prevalence of Type 3-like structures after the high Classical style and the prevalence of binary forms in the earlier part of the eighteenth-century. As I will discuss in Chapter V, this leads to questions of whether nineteenth-century movements with Type 2 27 Caplin, Classical Form, 173-174. Note, however, that Caplin does observe that this practice is “normative in midcentury works” (173). 28 Hepokoski and Darcy, Elements of Sonata Theory, 11. 25 characteristics share more than superficial features of their eighteenth-century predecessors—or whether the underlying processes and theoretical assumptions upon which they’re built are in fact entirely different. Throughout this dissertation, I also draw from L. Poundie Burstein’s Journeys through Galant Expositions.29 Burstein’s approach is rooted in the galant style. Its focus on trajectories between formal cadences accounts for some galant pieces in which some cadential punctuation that are ubiquitous in late-eighteenth-century works (especially medial caesuras) are absent or less defined. Analyzing this repertoire applying Burstein’s approach places it at the center of the conversation, viewing these pieces not only as predecessors or the works of “lesser” composers, but as material worth studying in its own right. Writing about the Type 2 sonata in the twentieth century has been all over the map. Descriptions and interpretations of the form can be grouped into several categories. The first is the “binary camp.” These theorists conceive of Type 2-like forms as essentially binary in nature, though they usually also give at least a nod to the form’s relationship to sonata form. It is in this group that “Scarlatti form,” as theorized by Ralph Kirkpatrick finds its home.30 It is so named because of Scarlatti’s frequent use of it and is also a type of Green’s balanced binary form (see Figure 1.2b). More generally, the present study is indebted to scholars who have treated mid- eighteenth-century galant music as a distinct, if related, style to those around it, including, 29 L. Poundie Burstein, Journeys through Galant Expositions (New York: Oxford University Press, 2020). 30 Ralph Kirkpatrick, Domenico Scarlatti (Princeton: Princeton University Press, 1953), 253-254. 26 but not limited to, Bathia Churgin, Mary Sue Morrow, and Daniel Heartz.31 While their work is not not necessarily focused on the Type 2 sonata, it’s a formal structure that comes up frequently in their discussions of the works of galant composers including Stamitz, Sammartini, and Boccherini (particularly those of Churgin and Morrow). Another view of the Type 2 sonata is that it is, in fact, essentially a Type 3 sonata with “reversed recapitulation.” This applies to sonatas in which the tonal resolution coincides with the return of S (and C, if present), which is then followed by P. Thus, the order of P and S from the exposition is seemingly reversed. This view tends to regard everything that follows the development as part of a “real” recapitulation, its own section on an equal structural level to the exposition and development. This group, which focuses mostly on nineteenth-century repertoire, includes Paul Wingfield and Timothy Jackson.32 I explore this view further in Chapter V. Finally, there are scholars who accept the Type 2 sonata as a valid analytical category and who have built on Hepokoski and Darcy’s research. Most notable among these is Peter H. Smith, who has provided the most direct case for the nineteenth-century Type 2 sonata.33 31 Bathia Churgin and Mary Sue Morrow, eds. The Eighteenth-Century Symphony, vol. 1 of The Symphonic Repertoire (Bloomington: Indiana University Press, 2012); Daniel Heartz, Music in European Capitals: The Galant Style (New York: W.W. Norton & Company, 2003). 32 Timothy L. Jackson, “The Tragic Reversed Recapitulation in the German Classical Tradition,” Journal of Music Theory 40, no. 1 (1996): 61–111. 33 Peter H. Smith, “The Type 2 Sonata in the Nineteenth Century: Two Case Studies from Mendelssohn and Dvořák,” Journal of Music Theory 63, no. 1 (2019): 103–138. 27 Outline of this Dissertation This dissertation explores three features of sonata form—the medial caesura, development, and the tonal resolution—in the context of Type 2 sonatas. The repertoire analyzed spans several decades from about 1740 to 1790. I focus in Chapters II and III on the symphonies of Johann Stamitz and the keyboard Johann Christian Bach, respectively; Type 2 sonatas constitute a large percentage of these composers’ sonata forms. In Chapter II, I analyze movements from two keyboard sonatas by J.C. Bach, with particular attention to the role of the medial caesura in shaping not just the structure of the exposition, but the impact of the tonal resolution in the next rotation. Chapter III explores the development and preparation of the tonal resolution, reinforcing how the development and tonal resolution are governed by a single rotation. Two symphonies by Johann Stamitz demonstrate the composer’s idiosyncratic treatment of rotation in his Type 2 forms. In Chapter IV, I analyze a keyboard sonata in E-flat major by Marianna D’Auenbrugg (published in 1780) and the first movement of Wolfgang Amadeus Mozart’s “Prague” symphony, K. 504 (composed in 1786); while D’Auenbrugg has only the one published sonata and Mozart’s sonata forms are overwhelmingly Type 3 sonatas, these are particularly striking examples of the expressive possibilities of the form, as well as its continued existence in the 1780s. In the final chapter, I look forward to the Type 2 sonata in the nineteenth century, where its viability as a distinct formal type remains under scrutiny. The musical examples in this dissertation can only scratch the surface of the behaviors and analytical issues of eighteenth-century Type 2 sonatas. The form is the framework for works with a wide variety of affects and contexts, and care should be taken 28 to avoid over-generalizing. Focusing on the works of specific composers provides a snapshot, a way in to additional studies of the Type 2 sonata. In this dissertation, my goal is to shed light on analytical issues related to the Type 2 sonata that may be reflected differently in other geographic regions (Italy, for instance), or that might be treated differently by other composers. In other words, the music described here is but a starting point, not intended to represent the entirety of the Type 2 sonata’s identity and behaviors. 29 CHAPTER II THE MEDIAL CAESURA The second movement of J.C. Bach’s piano sonata in D major W A2 is a demure follow- up to the sonata’s outgoing and declarative first movement. The light, two-voice texture of this “andante di molto” consists of a lyrical melody and a gentle, string-like accompaniment. The affect of the movement’s opening is serene and uncomplicated—an opening remark that doesn’t ruffle any feathers. If the first eight measures are like the start of a gentle address, it is in m. 8 that the speaker comes to the end of a thought and takes a breath. The breath represents both the end of the previous idea and the impending start of the next one. While this piece is not texted, there are several musical signals that evoke the sense of a breath taken. Example 2.1 shows mm. 8-9 of the movement. Example 2.1: J.C. Bach Keyboard Sonata W A2/ii, mm. 8-9. 30 Perhaps the most obvious signal is the full beat of rest in the melody: literal silence from our speaker. The melodic pause is preceded by a half cadence (HC) V6/4-5/3, with ^3 and ^2 in the soprano melody. This imparts a sense of arrival but not completion— there’s still more of this speech to come. Finally, the action resumes with new material, thematically distinct from the opening eight measures. The cumulative effect of all of these signals is a division of the exposition into two parts—in this case between a first and second theme. It is both a moment of closure and a moment of possibility, and it plays a defining role in the exposition’s formal design. Hepokoski and Darcy call the gap of m. 8 the medial caesura (MC). As in W A2/ii above, at the MC the division of the exposition into two parts occurs by means of three different (though interrelated) parameters:34 rhetoric (the pause itself), harmony (the HC that ends the tonic first part of the exposition and opens the door for the dominant second part), and theme (the distinction in thematic content between P/TR and S/C).35 Due to its role of dividing the first part of a sonata, the MC is one of the two most important cadences in an exposition (the other is the essential expositional closure, or EEC, the first satisfactory PAC in the new key that proceeds to new material).36 While Sonata Theory’s terminology originated in the twentieth century, L. Poundie Burstein 34 I follow Hepokoski and Darcy’s use of MC to apply not just to the caesura itself but to the whole event from attainment of half cadence, prolongation, and caesura. 35 S can be P-based; sometimes retrospective interpretation is needed to determine which action zone is in play. 36 Hepokoski and Darcy, Elements of Sonata Theory, 18. 31 links the concept of the MC to Heinrich Christoph Koch’s förmliche Absatz.37 It is normally a highly salient event, with features that are recognizable in the score and by ear. From a bird’s-eye, spatially-oriented view of a sonata, the MC divides the exposition into two parts, yet is also the culmination of a musical trajectory moving through time. In other words, the MC is not created after the fact; it is part of the sonata’s unfolding in time. We hear the MC when it happens, and hearing it orients us in sonata-space and sonata-time. Studying how an MC is articulated helps us better understand what comes both before and after it in a larger musical context. This chapter examines the role of the MC in Type 2 sonatas specifically. Since the MC is the musical event that makes S-space available, it sometimes plays an even greater role in the formal trajectories of Type 2 sonatas: just as the MC forms the seam between Parts 1 and 2 of the exposition, it is sometimes reproduced exactly (or exactly but for transposition) at the seam between development and recapitulation.38 When the HC effect is strong at this seam, it retains the role of MC—it is a caesura happening partway through the second rotation. The difference in a Type 2 sonata is that the MC occurs between the development and the tonal resolution, as opposed to a Type 3 sonata in which this caesura is followed by the “real” MC—the one familiar as the divider between parts 1 and 2 of the exposition. It is not necessary for the MC of the second rotation in a Type 2 sonata to match that of the exposition, or even to resemble it. The differences can 37 L. Poundie Burstein, Journeys through Galant Expositions (New York: Oxford University Press, 2020), 81. 38 Many developments end with a clear HC leading into the tonal development or resolution, though it also happens sometimes that the line between HC and PAC is blurred. L. Poundie Burstein, in “The Half Cadence and Other Such Slippery Events,” Music Theory Spectrum 36 no. 2 (2014): 203–227, discusses this seam and its apparent contradictions. 32 be instructive: one can ask how their characters compare—do the differences between the exposition’s MC point and the development-tonal resolution seam tell us something about the expressive trajectory of the sonata as a whole? One can also pay attention to how this affects the entrance of S—is there a sense of contrast, of anticipation? Is it highlighted more or less than in the first rotation? In placing both MCs in dialogue with one another—and situating them within the broader sonata narrative—we gain a better sense of how the two rotations relate to one another tonally and rhetorically. This in turn bolsters our view of an individual movement’s formal journey. In Bach’s W A2 (ii), the parallels between part 1 of the exposition (mm. 1-8) and the development are readily apparent. More often than not, however, the parallels between the MCs of rotations 1 and 2 are subtler, and sometimes such parallels are either very weak or absent all together. In such cases, we can consider if and how the passage leading up to the dominant pedal changes how we hear the onset of S. In other words, the MC is not merely a theoretical construct; it has real implications for the listener, analyst, and performer. As we listen to a piece in time, the MC is a formal landmark, understood in the context of many norms and expectations, that gives us information about where we are in the movement and what we can reasonably expect to happen next. Performers must decide how to prepare this moment, execute the half cadence and gap, and launch the next theme in a way that situates each element in the broader expressive narrative of the exposition (given the relatively small scale and light affect of this movement, and the fact that it becomes clear during the caesura that S must follow rather than P, the descending left hand pentascale at both formal seams can be laid-back and understated, a release of the energy accumulated via pedal points). More broadly, and 33 applicable to listeners, analysts, and performers alike, understanding the MC gives us a sense of how the parts of the exposition hang together. Throughout this chapter, I take J.C. Bach’s keyboard sonatas as a case study. I have chosen these works as my corpus because 1) Bach’s “cosmopolitan career” placed him in direct contact with other well-known Type 2 composers such as Sammartini in Milan and Carl Stamitz in Mannheim;39 2) Bach’s influence on the young Mozart is well- documented,40 but the relationship between their formal practices—especially in Type 2 forms—remains underexplored; and 3) the formal structures of Bach’s sonata movements are highly varied and make evident the range of expressive option available within the Type 2 paradigm. Perhaps most importantly, these are wonderful pieces that are too often appreciated as “proto-Mozartean” rather than appreciated in their own right. By focusing on Bach’s sonatas as they are, and not just in terms of their influence on Mozart, we have the opportunity to situate them in, and view them through the lens of, their own time. I begin by briefly examining MC norms as presented by Hepokoski and Darcy. After outlining harmonic and rhetorical expectations, I explore the MC in Bach’s keyboard sonatas. As I demonstrate, there is much from Hepokoski and Darcy’s theory that accurately describes Bach’s MC practices. I also show, however, that there are also significant differences between the formal expectations of Bach’s sonatas and those of his 39 Paul Corneilson, J.C. Bach, in The Late Eighteenth-Century Composers, series ed. Simon P. Keefe (New York: Routledge, 2016). 40 Perhaps the most overt musical connection between Bach and Mozart is the younger composer’s set of piano concertos (K. 107) based on the elder’s keyboard sonatas (W A2, W A3, and W A4). 34 predecessors. These differences are magnified by the role of Type 2 sonatas within this corpus, which make up almost half of the sonata-form movements in the keyboard sonatas. I proceed to a discussion of exposition types, which are differentiated from one another by their treatment of the MC. Hepokoski and Darcy identify two exposition types: the two-part exposition (which has one MC) and the continuous exposition (which does not have an MC). To this list I add a third, which Hepokoski and Darcy discuss in the context of S-complications: the exposition with “apparent double medial caesura effect.”41 All three exposition types are found in Bach’s keyboard sonatas, although the dominance of the two-part exposition evident in late-eighteenth-century works is considerably weaker. I show how exposition type interacts with a Type 2’s tonal resolution, focusing on the role of the crux in the second rotation—the point at which the rotation’s thematic trajectory locks into the exposition’s referential layout.42 I outline some of the primary strategies available at this key moment to set up the half cadence that leads to the return of S and how these strategies emanate from expositional structure. At the end of the chapter, I analyze four complete movements from Bach’s keyboard sonatas to show how the MC becomes central to a work’s “Type 2-ness,” and how its effect is dependent on the features around it. Each piece analyzed is in dialogue with a different exposition type. First, I return to the Keyboard Sonata in D major, W A2 (ii) to situate the MC as discussed at the start of the chapter within the trajectory of the 41 Hepokoski and Darcy, Elements of Sonata Theory, 23 (two-part exposition), 51 (continuous exposition), 170 (apparent double medial caesura). 42 Hepokoski and Darcy adapt this term from Ralph Kirkpatrick’s use of it in regard to Scarlatti’s sonata forms. See Hepokoski and Darcy, Elements of Sonata Theory, 239-241; Kirkpatrick, Domenico Scarlatti, 253-261. 35 movement as a whole. Next, I examine the Keyboard Sonata in B-flat major, W A1 (i), which has a two-part exposition and is fairly representative of Bach’s Type 2 sonatas. Third, I explore the Keyboard Sonata in C minor, W A6 (i), a highly unusual movement whose continuous expositional structure relates to the entire movement in equally fascinating ways. Finally, I end with an analysis of another keyboard sonata in C minor, A A8a (iii), which is in dialogue on some level with both the double medial caesura effect and the continuous exposition. I consider how various interpretations of important cadences affect our reading of the whole sonata’s structure. In all of these analyses, the Type 2 sonata is not just an architectonic structure: it has expressive implications as well. Medial Caesura Norms Hepokoski and Darcy were by no means the first to discuss this moment of formal articulation; they specifically cite similar concepts in Leonard Ratner’s and Karol Berger’s theories of sonata form.43 However, their comprehensive catalogue of MC characteristics and treatment of it as a central aspect of Sonata Theory is unprecedented. Mary Sue Morrow has written about the challenges of writing about the symphonic repertoire from c. 1700-1760, which she describes as a “period in flux”: “Many of the earliest binary/sonata movements can be understood in terms of Hepokoski 43 Leonard Ratner, Classic Music: Expression, Form, and Style, (New York: Schirmer, 1980), 207–247; Karol Berger, “The Second-Movement Punctuation Form in Mozart’s Piano Concertos: the Andantino of K. 449,” in Bericht über den Internationalen Mozart-Kongreß Salzburg 1991, Mozart-Jahrbuch 1991, ed. R. Angermüller, D. Berke, U. Hofmann, and W. Rehm (1992), 168–172. 36 and Darcy’s theory in general, though their default levels do not always apply.”44 One of the challenges of studying this music is to avoid overgeneralizing: musical styles of the eighteenth century are highly varied. That said, Sonata Theory works very well to describe much of this repertoire—or at the very least, to provide the foundations that can be modified or extended to describe it. This is certainly true of the MC. While the default options are different, there is much to be gained from reading these pieces through a Sonata Theory lens. A case in point: in the late-eighteenth-century sonata, by far the most common type of MC according to Hepokoski and Darcy, was a half cadence in I or V (represented as I:HC or V:HC, respectively) followed by a caesura. Small-scale movements were more likely to have a I:HC MC than a V:HC MC; the opposite is true for larger movements (this is by no means an inviolable rule. As movements grow in scale in the latter half of the eighteenth century, the V:HC MC option becomes comparably more common. The preferences of individual composers are likely subordinate to the scale of the movement in the evaluation of MC norms—the fact that Mozart’s keyboard sonatas more often employ V:HC MCs than J.C. Bach’s probably has more to do with the size of their respective movements than the fact that one is Mozart and one is Bach. Nevertheless, it is important to keep this trend in mind when considering MC norms in the mid- or late-eighteenth centuries. 44 Mary Sue Morrow, “Sketch for a History of the Eighteenth-Century Symphonic Repertoire,” in The Eighteenth-Century Symphony, vol. 1 of The Symphonic Repertoire, ed. Mary Sue Morrow and Bathia Churgin (Bloomington: Indiana University Press, 2012), 779–792. Quote 1 (“period in flux”) is from p. 780, different default options p. 783. Both are quoted in James Hepokoski, A Sonata Theory Handbook, (New York: Oxford University Press, 2020), 53. 37 Certain aspects of the MC remain essentially constant throughout the eighteenth century. Harmonically, the V:HC is stronger than the I:HC; however, in expositions with two MCs—in which the first MC is usually a I:HC and the second a V:HC—it is quite possible for the V:HC to be articulated more weakly than the I:HC. Weakening factors may be harmonic (for instance, minimal preparation of the dominant arrival) or rhetorical (for instance, little to no melodic separation across the implied gap that suggests continuity—a type of caesura fill).45 There are several musical signals of the half cadence’s arrival. The MC may be prepared by a dominant lock in which V is attained and then prolonged before the actual caesura gap. This prolongation can be brief, or it can be fairly extensive, lasting several measures. The cadence can also be approached by chromatic bass motion ^4-^#4-^5, which signals the tonicization of the dominant and intensifies its arrival. Perhaps most essential feature of MC preparation is energy gain, in which there is an increase of dynamics and rhythmic, and harmonic-rhythmic activity into the MC (this energy dissipates over the MC gap and S-space opens at this lower energy level). Energy often culminates in hammer blows, “ostentatious” reiterations of the dominant chord.46 These are characteristic around both the MC and the EEC (and the essential structural closure, or ESC—the EEC’s counterpart in the tonal resolution. The ESC is the first strong cadence in the home key after P in the tonal resolution) and can be found in countless eighteenth- 45 Hepokoski and Darcy, Elements of Sonata Theory, 163–171. 46 Hepokoski and Darcy, Elements of Sonata Theory, 36. 38 century movements. Dominant lock, chromatic bass motion, energy gain, and hammer blows are aurally salient and recognizable signals of the MC. Hepokoski and Darcy identify several common deformations and lower-level defaults for MC treatment in late-eighteenth-century repertoire.47 Of greatest relevance for mid-century works is what they call the “incipient medial caesura.” In these situations, there might be the suggestion of a half cadence without a clear launching of a new zone, or there might not be a satisfactory S theme following a clear half cadence. Hepokoski and Darcy associate the incipient MC with early sonatas of the 1740s, 1750s, and 1760s.48 In such a case it may be preferable to focus on how individual elements of the potential MC event are articulated or subdued, rather than trying to unequivocally label it as a “real” MC or part of a continuous exposition. J.C. Bach’s keyboard sonatas provide a valuable way into this topic for several reasons. First, ten out of the twenty-four sonata form movements in his keyboard sonatas are in dialogue with the Type 2 form (the other fourteen are Type 3 sonatas). There are no Type 1 sonatas. The formal practices represented in these movements are quite diverse and thus allow us to explore a number of practices within a single composers’ output (and within a single genre at that!) that are relevant to other composers’ works as well. 47 For an in-depth examination of MC deformations in Beethoven’s Type 3 sonata forms, see Mark Richards, “Beethoven and the Obscured Medial Caesura: A Study in the Transformation of Style,” Music Theory Spectrum 35 no.2 (2013): 166–193. 48 Hepokoski and Darcy, Elements of Sonata Theory, 63–64. 39 In Bach’s sonatas, the I:HC MC is far and away the most common HC type, but there are also a few notable instances of incipient or otherwise undermined MCs. Figure 2.1 shows Bach’s Type 2 keyboard sonata movements and their MC types. Figure 2.1: MC types in J.C. Bach’s Type 2 keyboard sonata movements Catalog Number49 (movement number) MC Type W A1 (i) I:HC W A2 (ii) I:HC W A3 (i) V:HC W A5 (i) I:HC W A6 (i) Continuous Exposition, subtype 2 (no MC) W A8a (ii) Continuous Exposition, subtype 1 W A9a (i) I:HC W A12 (i) I:HC W A12 (ii) I:HC, V:HC W A12 (iii) I:HC, V:HC W A18 (i) I:HC W A21 (i) V:HC About a third of Bach’s keyboard sonata two-part expositions have a V:HC MC. Nearly twice that number have a I:HC MC, and the remainder are either continuous expositions or expositions with a double medial caesura effect. Amongst the ten Type 2 sonata movements, nearly all have two-part expositions with a single MC. Five have I:HC MCs, two have V:HC MCs, another two are best read as continuous expositions (one each of Hepokoski and Darcy’s two subtypes), and two others have an apparent double medial caesura effect. These assignments are by no means clear-cut; there are cases, for instance, in which an MC could perhaps be interpreted based on harmonic content and 49 I use the catalog numbers found in The Collected Works of Johann Christian Bach 1735–1782, vol. 48, “Thematic Catalogue and Music Supplement,” compiled by Ernest Warburton (New York: Garland Publishing, 1999). 40 placement within the larger phrase rhythm but ultimately lack the sense of repose or division necessary for a sufficient MC effect. I explore this at the end of the chapter in an analysis of Bach’s Keyboard Sonata in C minor, W A8a (iii). Formal categorizations themselves require significant interpretation, and there is some room for different analytical outcomes. Only two movements with two-part expositions (W A3 (i) and W A21 (i)) have V:HC MCs; only one movement is best placed in dialogue with the continuous exposition (WA 6 (i)); and one (W A8a) maintains a participates in a dialogue between the continuous and double MC forms. Between I:HC and V:HCs—and between two-part and continuous expositions or double MC effects—Bach appears to have had a preference for the former in each case, and we can consider those options norms within this corpus. Knowing this, analyzing MCs can go in several directions; the most thorough, comprehensive view will consider all of the following factors as applicable. For the moment, we will assume a piece has an MC. First, one must consider its position in the overall tonality of the movement: is it a I:HC, or a V:HC (or its equivalent), or a rare instance of a lower-level default (such as a V:PAC)? Once the harmonic orientation of the cadence is noted, the question turns to the caesura and its preparation: is there a literal gap in sound? Do any voices bridge the gap sonically with caesura fill? How long is the gap (one might also consider how the gap is contextualized hypermetrically)? How is the caesura itself prepared (consider harmonic progression; scale degree motion; prolongation; thematic, motivic, or rhythmic manipulations; and energy gained before the caesura). Finally, we can turn to the launching of S itself: how does S compare with 41 the material that preceded it? Does it represent a sharp shift in affect? Does it borrow some motivic or textural characteristic from part 1? Once we have asked these questions and considered the rotation 1 MC through that lens, we can turn to rotation 2. As noted above, the material leading up to the return of S rarely corresponds exactly to that of the exposition. In content they may be different, but in the context of a Type 2 sonata, both serve the same purpose: to clear the way for the return of S. The same questions apply as in the exposition, with a few additions. First, consider how the material is different. Such a comparison takes a few forms aside from merely noting that the material is different. Consider its effect on energy gain, consider how it promotes or undermines a sense of rotation, consider whether there are any motives, gestures, or textures that recall the parallel moment in the exposition, even if that material doesn’t return, strictly speaking. Does this moment set the return of S up as a surprise, does it flow “logically” into S? How distinct are the sense of development and the full restatement of S from one another? The final step is to consider how these two moments and their relationship fit into the overall narrative of the sonata as a whole. To what extent are these moments pillars of the unfolding drama? Is there a sense of intensification between the two? How much is the onset of the tonal resolution emphasized? In what ways might the tonal resolution respond to the events of the exposition to form a cohesive whole? Exposition Types The following section explores the mechanics of three exposition types: two-part expositions, continuous expositions, and expositions with double MC effect. Since much 42 of the chapter thus far has been concerned with two-part expositions, this section focuses on the latter two types. Two-Part Expositions In Sonata Theory, exposition types are defined by the number of medial caesuras they contain. In other words, it is the punctuation provided by the MC that defines exposition structure, rather than, say, a specific configuration or number of themes. There are two basic exposition types in the eighteenth-century sonata: the two-part exposition and the continuous exposition. In a two-part exposition, there is one MC that divides the exposition into two parts; it separates P-TR (part 1) from S / C (part 2). In a continuous exposition, there is no MC (and thus there can be no S). It is also possible for a sonata exposition to have a double MC effect, in which a second MC divides S into two modules. In such cases, the second MC is less fundamental to the structure than the first. The overall structure, shown in Figure 2.2, is best defined as a two-part exposition in which the second part is also divided. Figure 2.2: Formal layout of a two-part exposition P TR ’ S C I MC V EEC Continuous Expositions According to Hepokoski and Darcy, it is the MC that “forcibly opens up S-space”; in other words, the MC not only signals the arrival of S, it is the event that makes the 43 launching of S possible in the first place.50 One of the oft-debated tenets of Sonata Theory is that “if there is no MC, there is no S.”51 The normative alternative is the continuous exposition, which can still have contrasting material between P and the rest of the exposition—it’s just not separated off by a cadence and thus lacks the punctuation that creates a division into two parts. Such a continuous exposition might still, in other words, have what, taken in isolation, might sound like an S theme; what Sonata Theory argues, and which I support, is the idea that the content of a theme is not enough to define formal function (after all, there is no single way to craft a P theme or an S theme)—context plays just as important a role in defining thematic identity.52 Continuous expositions have no medial caesura and thus consist of a single span from the start of P to the final cadence of the exposition. They ought not be defined unequivocally as lacking an MC, as though they are somehow less than fully-formed or as though something has been forcefully taken from them; they are complete, viable, and normative structures in eighteenth-century instrumental music. When I say, then, that continuous expositions have no S theme, it is not a judgment about value or even complexity so much as a point of comparison between them and the two-part expositions 50 Hepokoski and Darcy, Elements of Sonata Theory, 25. 51 Hepokoski and Darcy, Elements of Sonata Theory, 117. Also note that Hepokoski clarifies that “even though the MC + S guideline is an ultra-strong classical norm that should not be set aside lightly, it is not a rigid rule, nor should it be applied as one” (James Hepokoski, A Sonata Theory Handbook [New York: Oxford University Press, 2020], 65). 52 The concept of the continuous exposition, especially as defined by Hepokoski and Darcy, is not without its skeptics. This includes William Caplin, who maintains that whether or not there is an “S” depends on thematic content and characteristics rather than what came before it. See William E. Caplin, “Response to the Comments,” in Musical Form, Forms, Formenlehre: Three Methodological Reflections, ed. by Pieter Bergé (Leuven: Leuven University Press, 2010), 59–60. 44 we tend to know better. There are several ways in which one can break down different subtypes of continuous exposition. Perhaps the most general one is to find that blurry line between continuous expositions that proceed without any suggestion of S (i.e., those in which no expectation of a MC or S is ever presented) and those that do seem to suggest the preparation of an MC and S but then deviate from that path at some point. By making this distinction we can differentiate between expositional narratives that essentially fulfill our expectations as listeners and those which surprise us (through the lens of this specific mechanism). Hepokoski and Darcy also identify two subtypes of the continuous exposition: the continuous exposition with expansion section and the continuous exposition with early PAC. These subtypes can sometimes be distinguished from one another by whether there’s a suggestion of an MC and S or not, but more precisely the two are differentiated by how they come to a close. I will address the first briefly, as there is but one exposition among Bach’s sonatas clearly in dialogue with this form (W A8a in C minor), which I analyze later in this chapter. I then explore the second subtype in another of Bach’s sonatas. In the continuous exposition with expansion section, TR basically spins out past the point where there could conceivably be a viable MC. Formal proportion and harmonic stability are the main factors that determine where this point would be, and it’s often a point that you only know is past when you’ve already passed it. In other words, if you’re listening and find yourself in a phrase that is repeatedly driving toward a PAC in the new key, it’s almost certainly too late for an MC—if there isn’t a point before than that serves as a convincing MC, you’re likely dealing with a continuous exposition with expansion 45 section. Figure 3a shows the organization of this exposition type.53 This subtype can be further divided into those expositions that seem to approach an MC and S but never actually reach it. Figure 2.3: Formal layouts of continuous expositions subtypes 1 and 2 a: Formal layout of a continuous exposition subtype 1 P TR Fortspinnung----------------C I V EEC The second continuous exposition subtype is a bit more complicated: the continuous exposition with early PAC in the new key followed by (varied) reiterations of the cadence.54 In these situations, a PAC in the new key occurs before there has been any kind of MC effect. Figure 2.3b shows the formal layout of this exposition subtype: b: Formal Layout of a continuous exposition subtype 2 P TR C I early V:PAC “real” EEC Hepokoski and Darcy list four characteristics that signal this formal structure: 1) an early but emphatic PAC in the new key, 2) a brief phrase (often four measures or fewer) 53 Hepokoski and Darcy, Elements of Sonata Theory, 52. 54 Hepokoski and Darcy, Elements of Sonata Theory, 60. 46 following the PAC, 3) a “recapturing” (often in a highly varied form) of the preceding cadence, and 4) immediate repetition of the brief, post-PAC phrase.55 In something of an inversion of this situation, it is also possible to reject placing a movement in dialogue with the continuous exposition in favor of a less thematically- defined binary form. Especially in mid-century works, continuous expositions may put the Type 2 sonata into close dialogue with what Douglass Green defines as “balanced binary form.”56 Balanced binary forms are often associated with Scarlatti’s sonatas. The ends of parts one and two of the sonata have rhyming endings (which Hepokoski and Darcy— following Green—describe as A1+B2//C +B1, where letters represent thematic areas and superscripts key areas 1 and 2).57 Green’s conception of balanced binary is not rotational and it is vague enough to describe forms—especially early eighteenth-century forms— that are not particularly in dialogue with sonata form. If we were to translate this formula into rotational terms but preserve the rhyming ending, we could present it as A1+B2//An +B1, where n might represent any number of key areas visited over the course of the development. This vaguer formula is basically an over-simplified version of the kind of referential layouts Hepokoski and Darcy employ in their book. How does balanced binary relate especially to continuous expositions? It’s a bit of a “square is a rectangle” situation: Type 2 sonatas are a subset of possible balanced 55 Hepokoski and Darcy, Elements of Sonata Theory, 60–61. 56 Douglass M. Green, Form in Tonal Music: An Introduction to Analysis (New York: Holt, Rinehart and Winston, Inc., 1965), 76. 57 Green, Form in Tonal Music, 355. 47 binary organizational plans, but not all balanced binaries are Type 2 sonatas. First, note that the dialogic nature of Sonata Theory discourages absolute criteria for whether something is or isn’t a sonata. In addition to questions of rotation and key relationships, one might also factor in genre, time period, and geographic region. Thematic differentiation and scale may also play a part in determining whether or not to consider something a sonata form: a sixteen-measure minuet might follow something akin to a balanced binary form, but its genre and brevity make analysis in terms of sonata form a less-than-compelling interpretation. With all of this in mind, consider that the MC is one salient indicator of both scale and tonal-rhetorical differentiation (though the thematic material on both sides can be related). Without it, especially in early- to mid-century works, other defining features are needed to promote a sonata form interpretation. What else do continuous expositions mean in the context of Type 2 sonatas? The most obvious complication at first glance might be that, since Type 2 sonatas have tonal resolutions coinciding with S and continuous expositions by definition have no S, how is it possible to have a Type 2 sonata with continuous exposition? This is where the nuances of the Type 2 sonata’s defining characteristics become especially important. Essentially, remember that “tonal resolution at S” is a convenient but over-simplified way of defining the Type 2 paradigm. “S” is doing a lot of work here, not just for S-zones proper, but also for anything significantly past P(1), the first module of P in a multi-modular primary theme area. How to define “significantly past” is up for debate, and resists strict codification: I argue that, in special cases, P2 at the tonal resolution can signal a Type 2 layout (especially if P1 is particularly notable—if it has a fanfare-like quality or a clear, memorable motivic identity), though it is far more common for a later module of TR to 48 coincide with the tonal resolution. Whether pieces with an early crux—the point at which, having deviated from the exposition’s referential layout, the second rotation locks back in to that layout (normally S)— will depend on the content of a specific movement and its themes. The continuous exposition is particularly fertile ground for exploring some of the defining issues of the Type 2 sonata. These issues invoke not only Sonata Theory, but have clear ties to Caplin’s theory of formal functions as well. There are two elements to this discussion that parallel the conversations that have been had about Type 2 sonatas generally. The first concerns identity: how might considering Type 2s with continuous expositions through the lens of formal functions reveal why this designation is preferable to something else (either a binary form or Type 3 paradigm—one in which there is a sense of P having been omitted)? The other is more oriented toward behavior: in the case of a Type 2 sonata with continuous exposition, how is the seam between development and tonal resolution made manifest, and how does that relate to the comparable seam in the exposition? In some ways, while the issue of continuous expositions in Type 2 sonatas might seem more confusing at first, it is actually a good case study for the effectiveness of the Type 2 paradigm as a model (and why it makes more sense than a Type 3 one). This is something Hepokoski and Darcy observe.58 The question is: if there is no S, as in a continuous exposition, what would the second rotation of a Type 2 sonata look like? Where in the thematic rotation does the tonal resolution occur? While S is the typical 58 Hepokoski and Darcy, Elements of Sonata Theory, 381. 49 theme coinciding with the tonal resolution, the most important thing is that the tonal resolution coincides with a theme that occurs after P (especially P1). There are a couple of points to address in this hypothetical argument. First, this demonstrates clearly how “Type 2-ness” is not just defined as tonal resolution coinciding with S; that is not a sufficient definition to account for what differentiates this form from the Type 3 sonata. To me, this is a good theoretical case for rotation as the underlying foundation of sonata type. That this bears out in the repertoire can be confirmed with analysis. Rather, if it speaks to this model, this framework could be illuminating about the expressive relationships between events and trajectories in a given sonata movement. Double Medial Caesura Effect As L. Poundie Burstein notes, there is evidence from both the musical repertoire and the writings of eighteenth-century theorists (including those of Riepel, Marpurg, and Koch) that this exposition type was more normative than is suggested by Sonata Theory.59 I treat the resulting exposition type not as a result of S-complications as Hepokoski and Darcy do, but rather as a compositional option on the level of two-part and continuous expositions. One compromise between the Sonata Theory and Burstein models would be to think of expositions with a double medial caesura effect as a subset of two-part expositions and then recognize that there such expositions have implications that must be considered beyond the two-part exposition with one MC. As such, we can read the second MC as 59 Burstein, Journeys, 68–69. 50 existing within part 2 of a two-part exposition, but that it bears a different narrative than a sonata with a single PAC span from MC to EEC. In the mid-eighteenth-century, perhaps even more so than in the late-eighteenth-century, the double MC effect is not a deformation, or even a complication, but rather a norm of expositional structure. Conceiving, on the other hand, of a sonata with two, hierarchically-equal MCs represents a completely different theoretical framework. In this matter we see tension between “container” and “trajectory” based descriptions of sonata form. Dividing an exposition into two parts implies space: each part, separated by means of a cadence, is its own container. Burstein sees a trajectory between the start of a formal structure and its end at a cadence point (and such containers and trajectories can be nested). It is a tension that manifests in some other ways in the language of Sonata Theory: parts 1 and 2 of the exposition are representative of metaphorical containers, though the trajectory across the span from the start of a section to its MC is more goal-oriented. The two MCs involved in this exposition type are not the same in either content or hierarchical importance. Burstein connects this to Riepel’s ranking of the weights of different types of “punctuation,” as well as to Riepel’s recommendation that the same kind of punctuation not appear twice in a row in order to prevent redundancy.60 The eighteenth-century pieces considered in this project support the relevance of these guidelines to this repertoire. In expositions that have two half cadences that are viable MC points, they are likely to be I:HC and V:HC, in that order. This reflects the 60 Burstein, Journeys, 66. 51 harmonic trajectory of an exposition: the I:HC represents a closer tie to the tonic key— where we’re coming from—while the V:HC (or III:HC) represents a closer tie to the dominant or relative major. Six of J.C. Bach’s keyboard sonata movements are in dialogue to some degree with the double MC effect, though they are not in this exposition type. In each of these case, the first MC effect is a I:HC and the second is a V:HC (or III:HC, in the case of the finale of W A8a). Cadential weight and clarity varies across all exposition types, but multiple caesuras offer additional opportunities for ambiguity. Several different configurations are possible: Both MCs are crystal clear. The first MC is really clear, and there’s one or more subsequent moment that could be an MC but which is not unquestionably one. There is a relatively early moment that could be an MC, and a later one that is clearly an MC. There are multiple places where there could be MCs, but none of them are especially clear. I define clarity for these purposes as 1) a clear harmonic progression leading to V either in the tonic, dominant, or equivalent, 2) a break in texture in at least one voice (clarity increases the more voices participate in the caesura-gap), and 3) the launching of a satisfactory S theme that drives toward the EEC. In practice, this means that while it is not always an easy task to identify a piece’s exposition type, dialogue with the double MC effect opens up even more opportunities for ambiguity. It’s not merely a matter of slotting exposition types into a single box— especially where double MCs are concerned because of the extra potential for ambiguity. 52 Doing so would leave a significant number of pieces unaccounted for, and it would also require glossing over the rhetorical effects of formal seams. How does the double MC effect affect Type 2 sonatas? Theoretically, it means that there are two potential crux points: with MC 1 or MC 2. In practice, this gets a little fuzzier, because the material leading into the dominant preparation of the tonal resolution need not be the material of TR or S1. In J.C. Bach’s keyboard sonatas, for instance, this is almost never the case (although in a few movements there are motivic, rhythmic, or melodic similarities that definitely call back to that material—without directly quoting it). None of these exposition structures belong to a single formal type; as such they are not predictive of formal type either. But because the tonal resolution in the second rotation of a Type 2 sonata coincides with S (or in the case of a continuous exposition, Fortspinnung or post-early-EEC material), exposition structure has a salient effect on the tonal resolution. Below I examine some of the relationships between the MC in the exposition, and the seam between development and recapitulation in the second rotation. In some of Bach’s Type 2 sonatas—W A2, W A6, W A7, W A8a, and W A12— the exact material leading up to the exposition does not return in the passage leading to the close of the development but something about the rhetoric of the moment recalls the build to the MC in the exposition. In most cases this arises from rhythmic motives, related melodic gestures or contours, and similar textures. Such musical recollections emphasize the Type 2-ness of a movement by easing us back into the crux, the “moment of rejoining 53 the events of the expositional pattern after once having departed from them.”61 In other words, the crux becomes less of a momentary event and more of a process, slowly working our way out of the development and its fragmentary, modulatory features and back into more thematically and tonally stable territory. I also find this somewhat clarifying in terms of why the “double return” criterion for sonata form is an inadequate determinant of which forms are in dialogue with sonata form (Sonata Theory rejects this outright, but certainly it’s still a term that pops up here and there).62 In fact, it heightens our sense of the second rotation’s continuity across what is ultimately a somewhat porous boundary between development and tonal resolution. The articulation or obscuring of an MC comes in a variety of flavors that are found in both Type 2 and Type 3 sonatas. However, the way the MC unfolds in the exposition warrants additional consideration in Type 2 sonatas. For one thing, in a Type 2 sonata, there will only be one caesura to contend with, one which stands in for the exposition’s MC in the overall second rotation of the sonata and which also functions as the cadence that opens the gate to the tonal resolution. This is not to say that the music surrounding these points will be identical, but rather that their function of introducing S is the same in both rotations. In a Type 3 sonata (with a two-part exposition), these functions are divided amongst two different cadences: the one that leads into the recapitulation and the one that corresponds to the expositional MC. This suggests a 61 Hepokoski and Darcy, Elements of Sonata Theory, 239–241. 62 See James Webster, “Sonata Form,” New Grove Dictionary of Music and Musicians, 2nd ed., ed. Stanley Sadie and John Tyrrell (London: Macmillan, 2001), 23:688; and a Sonata Theory response in Hepokoski and Darcy, Elements of Sonata Theory, 343. 54 fundamentally different relationship between the function of the arrival at the development-tonal resolution of a Type 2 sonata compared with that of the onset of a recapitulation in a Type 3 sonata. This has implications for the trajectory of the entire movement, tied to its crux. In the case of an early crux, where the thematic material of the sonata’s part 2 locks into the track of the expositional rotation before S, in galant style both the drama and actual thematic content will parallel what occurred in the exposition (in other words, it will be no more or no less intense than its expositional counterpart. Many of Stamitz’s symphonies in particular have early cruxes, while it is a rare phenomenon in Bach’s sonatas. While MC treatment and exposition type are not predictive of sonata type, they do have different implications for the second rotations of Type 2 and Type 3 sonatas. The MC’s importance in the context of Type 2 sonatas stems from the fact that it is the MC that makes available S-space. In the second rotation of a Type 2 sonata, another half cadence is required to usher in S. The relationship between these two musical events need not be one of literal, or even transposed, repetition. In other words, the actual musical material just before the tonal resolution does not have to be the same as that which came before the MC in the exposition. In fact, at least in Bach’s keyboard sonatas, this is rarely the case. Essentially that means that the crux—the point at which the thematic tracking of the exposition’s referential layout settles for good—of Type 2 sonatas with two-part expositions can be reasonably expected to occur with S. This further reinforces the “clean” break between what is basically parts 1 and 2 of the second rotation: development and tonal resolution. 55 There’s a bit of a conundrum here, for all that these analogies are clarifying. In spite of these similarities, the divisions within these rotations are built on somewhat different premises. One of the successes of Sonata Theory is that it considers not just the tonal narrative of sonata form, but its thematic and rhetorical ones as well. Parts 1 and 2 of the exposition are certainly differentiated from one another by tonal and thematic means, but they are defined largely by their positions on either side of a rhetorical event. By talking about “development” and “tonal resolution” the emphasis is shifted away from the rhetorical punctuation that underlies the second rotation, just as in the first rotation. This is partly an issue of “form as container” versus “form as journey”: while it’s true that the musical spaces on either side of the half cadence are distinct in their function, and likely their content, focusing on this difference is not sufficient for a thorough understanding of a work’s formal layout—we also need to know how we pass between them.63 This is perhaps somewhat compensatory. Hepokoski and Darcy stress again and again that the return of S at the tonal resolution does not constitute a recapitulation, even though it is launching a new section within the second rotation.64 All of these things are true, but I argue that when it comes to the music itself, the division between development and the tonic return of S is just as much rhetorical as it is tonal. With that in mind, I now examine the tonal, rhetorical, and thematic angles of the MC: 63 See L. Poundie Burstein, Journeys, 5–10 and Mark Evan Bonds, “The Spatial Representation of Musical Form,” Journal of Musicology 27 no. 3 (2010): 265–303 on spatial versus temporal perspectives of form. 64 See, for instance, Hepokoski and Darcy, Elements of Sonata Theory, 353–355. 56 Tonal: Medial caesuras built around half cadences in either the tonic or the dominant reinforce the analogy between the bipartite structures of rotations 1 and 2. In situations with tonic half cadences, the analogy might be supported by the same harmonies—in any case the arrival point is the same. In the case of half cadences in the dominant, this is not the case: the tonal expectation established there is centered around the wrong axis (that of the dominant) for the tonal resolution. What is analogous here is the relationship between the tonal areas implied before and after the caesura. In the first rotation, we arrive at the dominant immediately before the caesura and remain there after it, whereas with a half cadence in the tonic, the tonality shifts following the gap. In sonatas in the minor mode, this shift is accentuated by the disjunction between the dominant of the original key and the launch of the second theme in the relative major. This is one more reason (in addition to their relative infrequency in galant sonatas) that sonatas in the minor mode can be so much more striking. In the second rotation, the dominant leads directly to its tonic, regardless of the exposition’s MC type. In short, the medial caesura and tonal resolution half cadences can be related through sharing the same harmonic content or by the tonal relationship across the caesura. The relationship between the expositional MC and the development’s terminal half cadence needn’t stop at the V chord that articulates each, however. Especially in the case of an early crux, the harmonic progressions (and thematic material they support) may run in parallel at the end of each of the respective phrases; indeed, this harmonic parallelism, intensifying by the end of the phrase, may be one signifier to the listener that we have locked back onto the exposition’s path. 57 Another consideration when comparing these moments is how strong the link feels between the tonal areas, something that also interacts with the rhetorical parameter discussed below. This ties into Burstein’s notion of the “slippery” half cadence.65 Generally speaking, the tonal pull across the seam between development and tonal resolution is likely going to be higher than in the exposition, where generally speaking there’s a close-parentheses effect to divide part 1 from part 2, while the end of the development leads in (i.e. moves forward into) the tonal resolution—even though “half cadences” are involved each time. Rhetorical: Some amount of syntactic punctuation (whether rhythmic, textural, or metric) participates in any half cadence.66 The silence of the MC separates parts 1 and 2 of the exposition. The separation is not just a matter of distance, however—S is launched (via the energy gained over the course of TR) from this point. The start of the caesura marks the end of part one, while the end of the caesura marks the start of part 2. The degree of separation between parts 1 and 2 of the exposition can vary dramatically—it’s not merely a question of “is a half cadence present?” or not. Relative rhetorical separation between the sections can be determined by a few factors: 1) the length of the silence, if there is one. A longer silence leads to more of a break than a short silence (or, of course, the absence of one). There are hypermetric issues that arise here 65 L. Poundie Burstein, “The Half Cadence and other such Slippery Events,” Music Theory Spectrum 36, no. 2 (Fall 2014): 203–227. 66 Burstein, “Half Cadence,” 206. 58 too: on what beat does the cadence conclude, and on what beat does S proper enter? How much of that time is silent? 2) How many voices participate in the caesura. This can be related to the first point. In other words, three voices may lead to a caesura gap, but if one carries over, the strength of the caesura effect is mitigated (note again, however, that this does not eliminate the caesura effect—it can still be implied if the majority of voices break. 3) Energy gain and contrast. The norm in the eighteenth-century sonata is for energy to build through the end of the P/TR block, that energy is released over the course of the caesura, and S begins at piano level. Energy gain manifests in terms of a) increased harmonic rhythm and b) increased surface rhythm.67 S is likely to begin in a calmer fashion, especially with slower harmonic rhythm, although Sonata Theory does not (rightly, in my view) define the distinction between P and S as one of melodic or thematic contrast. The start of S usually gives us a sense of the key (leading up to its confirmation with the EEC). Thematic: This is the most variable and hardest to predict among these parameters. Instead, what’s in play at the most basic level here is how we receive the thematic cue that we’ve entered reprise territory. There are two ways this can be accomplished: 1) by a clean thematic/motivic break between the end of the development/retransition and the start of S, or with a more gradual sense of tonal and thematic return, in which tonal resolution becomes more of a process than an event. 67 Hepokoski and Darcy, Elements of Sonata Theory, 94. 59 As case studies for each of these exposition structures and their effect on the tonal resolution, I analyze four movements from keyboard sonatas by J.C. Bach. I begin with the second movement of the Keyboard Sonata in D major, W A2. This movement has a two-part exposition and a close correspondence between the MC in the exposition and the seam between development and tonal resolution. The elements of the MC, including a clear example of caesura-fill, are well defined and salient. It is a much gentler break than its counterpart in the allegro first movement, the character of each movement represented in the span of a few beats. The Keyboard Sonata in B-flat major, W A1, also has a two-part exposition. In this case, however, the MC and development-tonal resolution seam are less similar than in W A2 (ii). Hypermetric and rhetorical differences amplify the caesura effect of the development-tonal resolution seam so that the start of S is completely set off from S. The Keyboard Sonata in C minor, W A6 (i) is a much more unusual movement. It has a continuous exposition, subtype 2 (early PAC in the new key), and similarities in the various themes across the PAC in the relative major set up a tonal resolution where the return to tonic is quite clear, but the corresponding thematic material is momentarily ambiguous. The final analysis, of the Keyboard Sonata in C minor, W A8a (iii) demonstrates the utility of a dialogic approach to expositional structure. Analysis of the exposition, which can be read in a couple of different ways, and resulting in possible readings of the exposition as continuous or as having a double MC effect, is essential to parsing out the elements of the development-tonal resolution seam. Interpretation of the large-scale form depends on how one hears these cadences and thus expositional structure as a whole. 60 Analysis: J.C. Bach, Keyboard Sonata in D major, W A2 (ii), two-part exposition Let’s return to the second movement of W A2, the example from the beginning of the chapter (see Example 2.1). Example 2.1: J.C. Bach, Keyboard Sonata in D major, W A2/ii, mm. 8-9. This MC is fairly typical of those found in eighteenth-century sonatas, though the division is not as strongly articulated as in other movements of its time. Consider, for instance, the MC of the first movement of the same sonata in Example 2.2: Example 2.2: J.C. Bach, Keyboard Sonata in D major, W A2/i, mm. 17-19 The main difference between the medial caesuras of the first and second movements is that the first movement (Example 2.2 has complete silence at its MC in m. 18, while in the second movement (Example 2.1) the bass fills in the m. 8 gap in the melody. The latter 61 case is an example of what Hepokoski and Darcy call caesura-fill, a common procedure in which a gap occurs in one or more voices while another one or more voices continue to sound through the gap.68 In other words, in m. 8 of W A2/ii, the fact that the last beat is filled in with sixteenth notes in the left hand does not override the rest in the melody. Those sixteenth notes are also an anacrusis into the start of S, so they group forward, as it were, to part 2 of the exposition rather than belonging to part 1. The first movement’s MC (returning to Example 2.2) differs not only in its complete break in sound in m. 18. It is in every way a more forceful point of arrival: there is a registral and textural crescendo in mm. 17 and 18 with the ascending arpeggios and four-voice texture on the downbeat of m. 18; the bass ^5 is struck three times in succession (which recalls the fanfare-like quality of the movement’s opening)—Sonata Theory calls these “hammer-blows”69—and the sharp textural contrast of the following m. 19. This, too, is a typical example of an MC, one more likely to be found in faster, more energetic first movements. Indeed, listening to these movements, the articulation of each movement’s MC reflects their respective characters. The gentler second movement takes a quieter approach, one that still implies a breath but which is less declarative. While there is no forcefulness around this transition, the V6/4-5/3 gesture sighing its way into the rest that follows, it is indeed this rhetorical point that sets the first part of this exposition (mm. 1–8) apart from the second (mm. 9–23). It is a landmark moment in this sonata specifically due 68 Hepokoski and Darcy, Elements of Sonata Theory, 40. 69 Hepokoski and Darcy, Elements of Sonata Theory, 34. 62 to the features mentioned above but also for the place it occupies in this norm of the genre: countless sonatas make use of this cadential divider (the alternative being a continuous exposition with no medial caesura, a viable, if less common compositional option). Having established how the MC operates within the context of the exposition, we turn now to the MC of the second rotation (which includes the tonal resolution). The dominant lock that lets us know the development is coming to an end begins in m. 33. Example 2.3 shows the retransition through the start of S in the tonic. Example 2.3: J.C. Bach, Keyboard Sonata in D major, W A2/ii, mm. 33-36 The melodic gestures leading to the MC in m. 35 are not exact transpositions of corresponding passage in the exposition (c. mm. 7-8), though the triplet on the downbeat of m. 35 recalls the triplets of mm. 5 and 7. There are a few similar elements, however: the quarter rest in the melody with caesura-fill in the accompaniment—and the descending pentascale in the accompaniment is the same gesture that introduced S in the exposition—as well as the pedal point anticipating the MC. The purpose of tracking these similarities is not to unlock the secret to predicting exactly when S will arrive in a given sonata, but rather to situate S’s arrival in time and retrospectively in the context of the 63 second rotation. It is through such contextualization that the birotational element of Type 2 sonatas is especially clear. Analysis: J.C. Bach, Keyboard Sonata in B-flat major, W A1/i, two-part exposition The first movement of Bach’s Keyboard Sonata in B-flat major, W A1 is a relatively straightforward example of a Type 2 sonata. It has a clearly divided two-part exposition and an unambiguous tonal resolution and second-rotation thematic return. Analyzing this movement, like the second movement of W A2, provides a point from which we can observe other ways in which a Type 2 sonata might unfold. It also shows how even a relatively simple Type 2 sonata has not just architectonic but expressive implications as well. Even between these two relatively straightforward examples of the Type 2 layout, the qualities of the MC within the larger form convey their contrasting characters. In W A1, the relationship between the two MCs is one of amplification, in which the MC’s function is stronger and clearer in the second rotation. In the exposition, a I:HC MC separates the exposition’s P and S themes. The P theme is clearly demarcated, eight measures in length, and followed by a series of musical afterthoughts (mm. 8-10 and its repetition in mm. 10-12, and mm. 12-14). These passages do not stand on their own: mm. 8-10 and 10-12 are reinforcing the preceding PAC in the tonic B-flat major (the entire exposition is extremely diatonic in the tonic and dominant tonal regions). Measures 12-14 drive toward the MC, a I:HC in m. 14. Example 2.4 shows the MC. 64 Example 2.4: J.C. Bach, Keyboard Sonata in B-flat major, W A1/i, mm. 10-14 The cadence is prepared IV-V6/4-V, arriving on ^2 in the soprano. The expected caesura occurs in the melody, while the single-line accompaniment continues with scalar caesura-fill. The harmonic component of the medial caesura is a half cadence in the tonic rather than the dominant—this is Bach’s overwhelming preference in the sonatas and characteristic of small-scale eighteenth-century movements more broadly. The medial caesura is the pause that allows us to shift gears between parts 1 and 2 of the exposition, but in this piece, its understated nature also permits a sense of continuity—almost like moving across a dotted line. The melodic material around the caesura is somewhat formulaic: scale degrees ^3-^2 in the soprano followed by caesura- fill—bass lead-in to the secondary theme that reaches up to scale degree 6 in V, then descending stepwise to the new tonic F major (see Example 4). This gesture, characteristic in expositions—indeed, Bach himself applies the same formula in other sonata movements—is not nearly as common leading into second-rotation tonal resolutions. While the first and second halves of a Type 2 sonata essentially run in thematic parallel with one another, the half cadence that prepares the tonal resolution need not be—and often is not—a literal or transposed iteration of the medial caesura as it occurred in the exposition. Even when these two moments are not the same, we can use them to 65 compare the degree of division between sections. Some analytical questions to ask when comparing these moments are 1) how clearly articulated is half cadence and tonal resolution event compared to the expositional MC event? and 2) does the context of one change something about how we hear S? Consider now how S is introduced in the second rotation and how that introduction compares to that of the first movement. First, the movement is clearly rotational: the relatively brief development begins with P in the dominant and continues with P-based content before dissolving into more gestural, retransitional materials. Example 2.5 shows the retransition and start of the tonal resolution. Example 2.5: J.C. Bach, Keyboard Sonata in B-flat major, W A1/i, mm. 54-67 The retransition includes a seven-measure dominant pedal culminating in arpeggiations of V. Though the entrances of S in both the exposition and tonal resolution are preceded 66 by half cadences in the tonic key, the return of S at the tonal resolution is heightened by not only this dominant prolongation, absent in the exposition, but also the metric shift of the caesura proper within the measure and by Bach’s fermata indication.70 The scale degree ^3-2 gesture and scalar bass motion of the exposition’s medial caesura does not return at all; primary theme material is bypassed altogether in “true” Type 2 fashion. Another difference from the exposition—there is a complete gap in sound between the half cadence and the start of S. While the pitch level remains constant across the gap, the complete break in sound compensates for that constancy, like a hard reset for the onset of S. The tonal arrival and caesura leading into the tonal resolution creates a sense of expectation greater than that of the medial caesura in the exposition. It is an amplification of the medial caesura’s initial invitation to the secondary theme: after a period of development and turmoil, we’re met not with the declarations and fanfare of the primary theme, but with the drive toward completion that only S can provide. The Type 2 narrative in this piece is not one of conflict, but rather one of amplification. In other words: there is no “problem” in the exposition’s structure that needs to be rectified, no conflict between competing elements in which one player emerges victorious. Instead, the second caesura point creates a more intense separation effect through metric and energy-level adjustments that heighten the feeling that the thematic return at the tonal resolution is initiating something. This is not to equate this moment with the onset of a recapitulation, rather the effect is more of creating distance between the process of 70 It is possible to use this fermata as an opportunity for a brief cadenza, which further widens the gap between the development section and tonal resolution. 67 development and thematic return. P and S are not the same in the exposition, but neither is there are shocking contrast of character between them. There is, however, a striking difference between the rhetoric of the development and that of the tonal resolution. The features mentioned above all play a role in accentuating that difference. In many ways this is another fairly typical example of a mid-eighteenth-century Type 2 sonata: it has a clear, two-part expositional structure with a first-level default (within Bach’s works) MC, an overtly rotational development, and a tonal resolution launched by S. This piece demonstrates but one way—amplification—in which the MC and tonal resolution can be in dialogue with one another. The following analyses explore Type 2 sonatas that, in part through their expositional structures, convey contrasting narratives. Analysis: J.C. Bach, Keyboard Sonata in C minor, W A6/i, continuous exposition J.C. Bach’s Sonata in C minor, W A6 (Op. 5, no. 6) is a curious work unlike others of its kind in Bach’s oeuvre. The Type 2 first movement gives way directly to a fugue, followed by a rondo with stylistic traits of a gavotte. None of Bach’s other keyboard sonatas resemble this structure in the slightest. The overall multi-movement organization of the piece could perhaps place this movement in dialogue with the overture (especially given its lack of repeats): a grand opening followed by a fugue, both preceding a dance 68 movement (in this case a gavotte).71 The effect of the first movement’s Type 2 structure is also unusual—especially among Bach’s sonata forms—for its handling of thematic return. The nature of the movement’s continuous exposition—it is one of only a few movements among the keyboard sonatas in dialogue at all with the continuous exposition—sets the stage for not only which materials are available at the tonal resolution, but also of how our expectations are primed at that crucial moment. Example 2.6 shows the sonata’s exposition. The exposition’s tonal trajectory takes us from the tonic C minor to the relative E- flat major. There is a i: PAC in m. 7, which ends P and launches a dissolving TR based on the same material. What at first seems like merely an inflection toward the relative major will ultimately retain its hold and, rather than produce a HC and MC, culminated in a III:PAC in m. 17 (the effect is mitigated both here and with the PAC in m. 7 by the sixteenth-rest delay of the root in the bass. Nevertheless, to my ear the PAC effect in both cases still stands). We have thus made it from the tonic to what is essentially the EEC in the relative major without an MC—there is neither a caesura or cadence that could represent it. Following the early EEC there is a “post-EEC” theme in the relative major that further confirms the new key with short phrases that do not bear the same functional weight as a proper thematic zone. The Type 2 portion of the sonata is best analyzed as a Type 2 form with a continuous exposition (subtype 2). In other words, there is no medial caesura in the 71 In Bach’s sonatas, the lack of repeats is an anomaly, though this lack was a characteristic of Johann Stamitz’s symphonies. 69 exposition; instead an early PAC in the new key (the relative major, III). The tonal resolution in the movement’s second rotation coincides with the return of the material that follows this early PAC. Figure 2.4 shows the complete sonata form. Example 2.6: J.C. Bach, Keyboard Sonata in C minor, W A6/i, mm. 1-19 70 Figure 2.4: Formal Layout of J.C. Bach, Keyboard Sonata in C minor, W A6/i Zone: P TR cadential repetitions (x2) Key: i early V:PAC “real” EEC Measure: 1 7 *17 *29 In this movement, the continuous expositional structure reinforces the continuity of textures and melodic gestures between TR and the post-EEC theme. The punctuation offered by the early PAC in the new key (making this a continuous exposition subtype 2) creates a rhetorical distinction between these thematic zones, but thematically, the post- EEC theme is almost an immediate continuation of part 1 of the exposition. My analysis shows how the continuous exposition with early PAC in the new key precludes a proper S theme; instead the exposition continues with a theme closely related to P. At the tonal resolution, these similarities briefly make it unclear which theme has returned, as though the themes have been synthesized into a single entity.72 I will begin this analysis with an exploration of the exposition; I argue that what happens in the exposition is essential to the expressive impact of the tonal resolution. Specifically, the combination of an early EEC (instead of an MC) and related thematic material on both sides of the cadence contribute to the perception of a real, P-based recapitulation in the second rotation that must then be retrospectively reinterpreted as a reprise of post-EEC material. 72 Hepokoski and Darcy, Elements of Sonata Theory, 258. 71 The source of much of this expressivity comes from that early EEC just over halfway through the exposition. The cadence in question is a PAC (perfect authentic cadence) in the relative major (III of the tonic C minor): E-flat major (m. 17). This is somewhat confusing, and disrupts our sense of where we are in the trajectory of the exposition. Transitional material precedes this cadence (mm. 7-17); characteristics that suggest we have entered TR include: 1) a dissolving statement of P (Hepokoski and Darcy would call this a “dissolving TR”),73 2) a shift away from the tonic C minor and the tonicization of E-flat major, and 3) harmonic sequences supporting new melodic material. Because we experience this material in time as TR, we can reasonably expect an MC at this point, likely a III: HC (a half cadence in the tonal area of III in reference to the tonic) given the strength of the E-flat tonicization within TR. This is not what Bach gives us, however. Instead, our point of arrival in m. 17 is a III: PAC, an emphatic confirmation of the new key. Hepokoski and Darcy consider the III: PAC to be a third-level default for an MC,74 so at first this seems plausible—certainly it occurs in the right spot in the exposition to serve as the MC. Upon further consideration, however, this reading becomes less satisfying for two main reasons: first, the material that follows is very short (only four measures plus an anacrusis—keep this in mind when we return to this theme in the second rotation—mm. 18-21). Second, it is fairly static and straightforward compared to the primary theme: it seems to have little other purpose than to complete the cadence; it is not a satisfying theme in and of itself. 73 Hepokoski and Darcy, Elements of Sonata Theory, 101–112. 74 Hepokoski and Darcy, Elements of Sonata Theory, 26. 72 Interpreting this cadence as the EEC better accounts for not only the cadence’s role in the exposition, but the function of the materials around it. The implications of such a reading would be that there is no S-theme proper, and that everything that follows is closing material. It occurs quite early in the exposition, however, such that nearly half of the exposition would constitute closing material. Hepokoski and Darcy’s four identifiers of the continuous exposition, subtype 2 are present, however: the early, emphatic PAC (m. 17), a brief phrase following the PAC (mm. 18-21), affirmation of the preceding cadence (m. 21), and repetition of the post-PAC phrase (mm. 22-25). While this interpretation may seem cumbersome, it actually accounts for the overall trajectory of the exposition and the brevity of what would otherwise be our S- theme. We can thus summarize the formal plan of the exposition as follows in Figure 2.5: Figure 2.5: Formal plan of J.C. Bach, Keyboard Sonata in C minor, W A6/i, exposition P--------TR----------early EEC----------EEC def.------ “Real” EEC m.1 m. 8 m. 17 m. 25 m.29 Because it occurs so early in the exposition, the PAC of m. 17 is a premature EEC that launches repetitions to reopen “seemingly closed authentic cadences.” This space can be reopened again and again until the “real” EEC provides sufficient closure to end the exposition.75 75 Hepokoski and Darcy, Elements of Sonata Theory, 60. 73 There is a degree of tension between the thematic layout outlined above and the specific motivic content of the exposition’s material, and the premature EEC (henceforth “EEC”) is its source. Because the early “EEC” is so clearly demarcated, the sense of arrival and of the completion of a process is strong. Additionally, the metric emphases change around the “EEC”: P has a sarabande-like quality with emphasis in the right hand on the second beat of the first three measures; after the “EEC” the anacrusis leads to strong downbeats in mm. 18-21. But there is also continuity between the pre- and post- “EEC” material; namely, the left-hand arpeggios that pervade P and TR are also characteristic of the material that follows the “EEC.” Bach obscures this somewhat by abbreviating the arpeggiations to the last three sixteenth notes of the measure during the sequences in TR (mm. 12-14), such that, while the post-“EEC” material clearly recalls P it also feels like a new process is beginning when the arpeggios resume. This tension between continuity and cadential punctuation becomes one of the most important features of the tonal resolution. The exposition is not complete, however: we are still searching for the “real” EEC. There is a real opportunity for closure after the repetition of the four-measure post- “EEC” phrase, but Bach withholds this from us with a stunning deceptive cadence on the flatted submediant C-flat (m. 25)! The material that follows (mm. 25-29) is new, harmonically unstable, and passes through the minor subdominant of E-flat (A-flat minor). This passage does not recapture the E-flat major cadence, rather, it defers the EEC, reopening the TR process and denying us closure. The drama of the deceptive cadence eventually dissipates, and we finally get our “real” EEC in m. 29—a release from the tension accumulated in m. 25. The exposition does not repeat—there is a sense of 74 diminuendo leading into the EEC from which a return to P would feel entirely incongruous. Instead we proceed directly into the development (mm. 30-48). While, given the dominance of the two-part exposition (and even, to a lesser extent, the continuous exposition subtype 1), this formal layout seems complicated it is not highly deformational. Bach uses sequential harmonic motion (exploiting the relative relationship between the exposition’s key areas) and similar motivic material such that there is a sense of progression and of continuity—in other words, the exposition’s layout feels natural. In fact, the deceptive cadence of m. 25 is the most unexpected element of the exposition, and that cadence is not specific to the continuous exposition structure. At this point, a listener can expect the development to begin with P material in the relative major, and for the tonal resolution to coincide with either the return of P or post-EEC material (at this point—the end of the exposition—it is too soon to determine sonata type). The rotational nature of the development is obscured slightly with the introduction of new material, but elements of P and TR peek through (consider the arpeggiating accompaniment in m. 35 and the ^3-^2-^1 approach into the cadence in m. 39). The retransition begins—perhaps indisputably—in m. 43 with the characteristic dominant-lock. Left-hand arpeggios alternate with throbbing unisons, heightening the anticipation of what is to come while also suspending us in time. The spell breaks in m. 47 with a descending pentascale ^5-^4-^3-^2-^1, delaying the arrival of ^1 by a sixteenth- note—the fulfillment of our expectation is delayed yet again, if only for a moment! Example 2.7 shows the retransition and the tonal resolution. 75 Example 2.7: J.C. Bach, Keyboard Sonata in C minor, W A6/i, mm. 40-56 Now we come to the tonal resolution. Here we might expect to learn whether we are dealing with a Type 2 or a Type 3 sonata.76 The Type 2 structure of the form is revealed in a surprising and wonderfully expressive way. Rather than clarifying the sonata type immediately, the tonal resolution begins with muddied waters. A C minor chord in the right hand fills the silence left by the delay of the left-hand arpeggio: this is the beginning of P! Because the arpeggiating accompaniment is held in common between P and the pseudo-S/C theme that follows the “EEC,” upon listening for the first time one has every 76 In some cases this is predicted at the start of the development; other times—as in this case—the sonata type is revealed later in the second rotation. See Hepokoski and Darcy, Elements of Sonata Theory, 369. 76 reason to expect a full recapitulation—and thus a Type 3 form overall. This expectation is allowed to stand until the anacrusis of the post-“EEC” theme disrupts this trajectory. This is surprising not only because P is set up so perfectly and convincingly, but also because the anacrusis emerges as part of a process that is just beginning, rather than out of one that has already ended, as it did in the exposition. The repetition of the four- measure phrase serves a very practical purpose in the second rotation—it allows the listener time to adjust their expectations.77 The rest of the second rotation plays out in much the same way as the first—at least until its conclusion. For Bach expands the material after the deceptive cadence (m. 56, now diatonic) and continues the unstable, developmental rhetoric that characterized the end of the exposition. Bach provides directions to improvise after striking a cadential 6/4 chord and then proceed directly into the fugue that follows. As a result, the ESC is elided with the start of the fugue, blurring yet another formal boundary through ambiguity. The tonal resolution portion of this sonata is relatively short, if striking. Its impact is directly related to the events and rhetoric of the exposition: all of the “Type 2-ness” of the second rotation is meaningful because it emerges from the ambiguity of formal function and the delaying of resolution in the exposition. By leading directly into a fugue, the sense of events unfolding step by step lends a stream of consciousness quality to this sonata that is unparalleled in Bach’s other works for solo keyboard. 77 This change of expectations (coupled with the start of the non-P-based development) is in dialogue with what Hepokoski and Darcy would call a Type 3=>Type 2 conversion. For a discussion of this phenomenon, please see Chapter IV. 77 These are not purely analytical matters; they have direct implications for performance as well. Perhaps the main question that arises from this analysis for the performer concerns cadences: to what degree do you anticipate an approaching point of arrival? Do you distinguish agogically between authentic cadences and deceptive ones (especially the chromatic one in m. 25)? And what do you do with the moment of tonal resolution? I do not believe there is a single answer to these questions, and I can imagine a performer approaching the piece differently every time. But I do believe that one cannot approach these moments in isolation, whether as an analyst or performer: each represents a link in a chain that leads us on, one step at a time. Analysis: J.C. Bach, Keyboard Sonata in C minor, W A8a/iii Another of Bach’s keyboard sonatas in C minor, the finale of W A8a, has an even more complex formal structure that plays with ambiguity and our sense of formal function over time and space. R.M. Longyear briefly analyzes this piece and includes it as an example of “binary sonata form.”78 While none of Bach’s Type 2 keyboard sonata movements clearly utilize a double medial caesura type (though it can be found in the symphonies and sonatas of Stamitz and Mozart), this particular movement is nonetheless in dialogue with that paradigm. Furthermore, it provides an example of the medial caesura declined and its potential role in Type 2 sonatas. Retrospective reinterpretation is not merely possible 78 R.M. Longyear, “Binary Variants of Early Classic Sonata Form,” Journal of Music Theory 13, no. 2 (1969): 174. 78 in this movement, I argue that it is essential to understanding how the piece’s tonal and thematic trajectories interact with one another. Example 2.8: J.C. Bach, Keyboard Sonata in C minor, W A8a (iii), mm. 1-13 The movement begins with a short four-measure phrase in which the melody rises from ^3-^4-^5. Example 2.8 shows P. The melody goes even a step further to ^5, leaping down to raised ^7; the combination of the short phrase length and rather anticlimactic cadence following the seemingly inexorable rise of the melody contribute to a sense that, while we have indeed attained a i:PAC, this is but the first module of P and there is more to come. The next four measures pick up on the lower-neighbor triplet motive of m. 3 and end 79 rather inconclusively on ^5. The final five measures of P take the energy of the first phrase’s ascent and channel them through descending chromatic motion from ^8 to ^5. At this point, we get a dramatic half cadence in m. 13. The dominant is attained through the descending chromatic motion of mm. 8-10, with several arpeggiations of V in the right hand. The action is intensified through octave doubling between the hands, arpeggiation through the octave, and nearly two beats between the sounding of the unison ^5 (V implied) and the next sounding pitch at the end of the measure. In terms of the size of most of Bach’s sonatas, m. 13 would be a very reasonable spot for an MC, especially given that P consists of several smaller subphrases and contrasting gestures. And indeed, this cadence meets several of the criteria for an MC: there is energy gain, a dominant arrival, and a distinct gap separating the material before and after the caesura. The material that follows his cadence, however, undermines this interpretation. While in the moment it might be possible to hear this cadence as the MC, the structure of the exposition in its entirety calls for retrospective reinterpretation to untangle the various strands of tonal, thematic, and rhetorical events that unfold over the exposition’s forty- seven measures. The theme that begins in m. 14, immediately after the i:HC, is a variation on P in the relative major. This version has a call-and-response quality: between each measure corresponding to P in i, another is inserted, based on the lower-neighbor triplet motive of P. Example 2.9 shows the beginning of this passage, from m. 14. This expansion of P goes beyond merely inserting a measure between each of the original gestures, it also doesn’t end with a PAC. 80 Example 2.9: J.C. Bach, Keyboard Sonata in C minor, W A8a/iii, mm. 14-23 Instead, it takes the lower-neighbor triplet motive in much the same way it was used in the second phrase of P (mm. 4 ff.) and runs with it until m. 23, a dominant arrival in the new key. There are a few conflicting factors to consider when determining how the theme beginning at m. 14 fits into the larger structure of the exposition. That it is so firmly grounded in III and that it begins after a HC and caesura could suggest the possibility of hearing this cadence as the MC (note also that there will not be another HC this clear in the remainder of the exposition). However, the content of this passage is based closely on P that feels like it is finishing a thought begun at the start of the movement. For this reason, I am inclined to hear the theme in m. 14 as TR—and as what Sonata Theory would call the grand consequent to P, which in retrospect forms a complex grand antecedent. The grand antecedent as a type of P consisting of “a lengthy, often 81 multimodular antecedent phrase”; in other words, it is made up of more than one subphrase or thematic module and ends with a I:HC (including sometimes with MC-like characteristics). It is often followed by a parallel grand consequent that begins with P (and which often functions as TR). A complex grand antecedent contains an interior cadence (in the case of W A8a, this would be the PAC in m. 4).79 While the complex grand antecedent works well to describe what’s happening thematically and rhetorically in W A8a, the tonal situation presents a bit of a problem: the phrase beginning in m. 14 is in III, where normally the grand consequent would present P material in the tonic key. It is a direct modulation; the gap is the only preparation for the new key. I have chosen to identify m. 14 as TR first because of the norm of the grand antecedent, because it varies P in phrase-structurally important ways, and because it dissolves into explicitly TR rhetoric in m. 22. Still, there remains some room for interpretation. Hepokoski and Darcy note, for instance, that it can be difficult to distinguish between the HC of a grand antecedent and an MC declined—a situation in which the material following a possible MC point fails to initiate part 2 of the exposition either tonally (by remaining in the tonic), thematically (by returning to P material), or rhetorically (by “reinvigorating” TR).80 In the case of W A8a, one might argue that the half cadence in m. 13 is the MC but it does not lead to an ultimately satisfying S theme. However, the relationship between the opening of P and what I call TR is best described as an antecedent-consequent relationship. 79 Hepokoski and Darcy, Elements of Sonata Theory, 77. 80 Hepokoski and Darcy, Elements of Sonata Theory, 45–47. 82 In my reading of the piece, we are still in part 1 of the exposition. The final measure of TR, m. 22 intensifies the rhythmic action with a steady stream of eighth-note triplets and rising chromatic motion ^3-^4-^5 in the bass. Example 2.10 shows this passage. Example 2.10: J.C. Bach, Keyboard Sonata in C minor, W A8a (iii), mm. 22-29 These are classic characteristics of the MC. So far, we have not actually had a cadence— the dominant of m. 22 is the start of the V chord that we expect to ultimately culminate in at V:HC. Still, it finally seems like this will be the “real” MC. It arrives a little late in the exposition for a Bach sonata, however, especially as the dominant is prolonged without generating a caesura effect. Broken-chord filler material in the treble is punctuated every two measures with a low ^5 (the second time in octaves) in the bass. The expected cadence and caesura never fully materialize. Both voices’ broken chord figures and dominant pedal carry across the point that we would expect the caesura to occur and what at first seemed to be the end of the gesture is stretched out over ten more measures. The last few measures of Example 10 show this passage and how it 83 continues, in a sense dissolving—it is as though the energy generated by the dominant arrival and filler material is all but spent before it manages to produce a satisfying MC. The lower-neighbor triplet motive returns in m. 31, and it is with that gesture that we push forward to the EEC. Closing material follows. Once we have come to the end of the exposition and taken into account all the various themes and cadences involved, the question of expositional structure remains. There is no single paradigm that adequately describes the form of this piece; rather, it is in dialogue with three different exposition types: the continuous exposition subtype 1 (the expansion-section subtype), the two-part exposition, and an exposition with double MC effect. At first this might seem contradictory—how can one piece seemingly have no MC or two MC effects? While no MC really materializes (placing the piece in dialogue with the continuous exposition), approaches are arguably made once or twice according to the above interpretations of the i:HC in m. 14 (the two-part exposition reading takes the grand antecedent reading and m. 27 as the MC, the double medial caesura effect reading treats the cadences in both mm. 14 and 27 as MC). To some extent, the question is not whether there are definitively any MCs in the exposition or not (and if so, how many), but rather how these moments relate to the norms of the genre and time period as well as to this particular movement’s structure. How one reads the exposition will have implications for how one understands the second rotation’s unfolding. Figure 2.6 shows possible readings of the first rotation. 84 Figure 2.6: J.C. Bach, Keyboard Sonata in C minor, W A8a/iii, possible readings of Rotation 1 a: J.C. Bach, Keyboard Sonata in C minor, W A8a/iii “grand antecedent” reading ROTATION 1 Antecedent Consequent m. 13 m. 14 I:HC III:HC P----------------TR----------------------------------------EEC i III b: MCs declined reading ROTATION 1 [MC1] [MC2] m. 13 m. 23 I:HC III:HC P----------------S1--------------------S2--------------------EEC i III There is so much going on in the exposition it is unclear how exactly the second rotation will play out. The main question is: where is the crux? Does it coincide with one of the moments discussed above? Is the articulation of each the same or different from the original versions? The rotation begins as expected, with P in III (in its original form from m. 1, not its P-based TR form). Figure 2.7 shows the layout of the second rotation corresponding to the exposition readings of Example 2.5 above. 85 Figure 2.7: J.C. Bach, Sonata in C minor, W A8a/iii, Rotation 2 a: J.C. Bach, Keyboard Sonata in C minor, W A8a/iii, formal layout of Rotation 2 ROTATION 2 Development [MC1] [MC2] m. 65 m. 82 v:HC i:HC P----------------S1--------------------S2---------------------ESC III v i b: tonal and thematic resolution occur together, continuous exposition version ROTATION 2 Development Tonal Resolution Ant. Cons. m. 65 m. 82 v:HC i:HC P--------------TR--------------------S2---------------------ESC III v i c: double MC reading with “wrong key” thematic return post-development ROTATION 2 Development “wrong key” thematic return tonal resolution m. 65 m. 82 v:HC i:HC P--------------------S1-------------------------S2---------------------ESC III v i 86 d: continuous exposition reading with “wrong key” thematic return post-development ROTATION 2 Development “wrong key” thematic return tonal resolution m. 65 m. 82 v:HC i:HC P--------------------TR--------------------S2---------------------ESC III v i All three P modules are heard in order—the first (corresponding to m. 1) in m. 48 at the start of the development, the second (corresponding to m. 5) in m. 52, and the third (corresponding to m. 8) in mm. 55. The first two are sounded in the dominant, and the third module explores other keys. It is repeated three times, first in F minor, then twice in G minor. The first two attempts are truncated and so do not end with a cadence. The third finally breaks free of whatever force was holding it back and leads to a v: HC, an exact transposition of the material from the exposition. When TR begins immediately after this cadence, as in the exposition, it begins in v. It is possible to imagine a different version of this piece, in which the sequences of mm. 55 and following lead to the dominant of C minor and set up a tonic-key return (see Figures 2.6c and 2.6d). This would effectively put the second-rotation TR at the point of tonal return. This would soften the Type 2 effect, especially because the material is so similar to the opening P. As it stands, however, the tonal work of the movement is not yet finished. In m. 69, Bach starts to modify the thematic material to produce the required modulation. He abandons the lower-neighbor triplet figure and passes through C minor 87 (a preview of the imminent tonal return) and F minor before landing on a striking V chord in m. 78. This V chord is the movement’s crux—it is the place where, having strayed from the exposition’s track in mm. 69-75, it locks back onto the exposition’s trajectory. Further modifications are minor and do not affect the overall layout of the rotation. The V chord corresponds to the V/III in m. 23. It is approached differently—with oscillating thirds and a stepwise lead-in at m. 77 rather than strictly stepwise motion in m. 22. The sense of anticipation is heightened here as we await the tonal return as well as by its role as crux: everything locks into place here, making the moment particularly dramatic. Stepping back and looking at the second rotation as a whole, how does it map on to the exposition’s referential layout? There is little new content, and those places tend to replace a relatively non-essential something from the exposition (as in m. 69 from m. 18). Several questions arise: is there a single moment at which we can say the tonal resolution occurs? Is that moment in line with the “recapitulation” of thematic material? What does this tell us about the two not-quite-MCs? With these questions too there are a couple of possible readings. One of the tricky things about deciding where development ends and tonal resolution stops is that our two possible MCs do not really produce the effect of launching something new. The first at least has a caesura to separate the before and after, but the latter does not—and that is where we finally reach the dominant of C minor. One possible reading hears a disconnect between the tonal resolution and the end of the development. In other words, the development ends in m. 64, and post- developmental space begins in the “wrong” key (v). As material leading into the MC declined is rewritten, this requires hearing 69-78 as what Rosen might call a “second 88 development,” in which material within a recapitulation takes on developmental qualities81). In W A8a, this takes a fair bit of time in proportion to the rest of the movement and, given that we are still in a foreign key, this reading is less convincing than its alternatives. If one hears P in the exposition as grand antecedent, this cadence is confined within part 1 of the exposition and belongs, as a unit, with its consequent (rather than breaking them up over a major formal event). Another possible reading hears the G minor statement of TR as part of the end of the development. In this case, tonal and thematic parameters are better aligned. It is also a simpler reading and compatible with P-as-grand-antecedent. What it doesn’t account for is the strong sense of arrival at the cadence in 64 and return of thematic material in m. 66. The sense of thematic return is heightened by the fact that we are coming off material (mm. 55-59) that had deviated from the exposition and now, for a few measures at least, we are back on track. All of this leaves the question of where the tonal resolution occurs. Neither of the arrival points involved usher in the new key at a specific moment in time—the first is strongly in G minor, and the second never breaks the train of thought as needed for an MC. It would be less productive to try to find an exact moment of tonal resolution than to outline the process by which it unfolds. This process begins in m. 78, just before the V chord and continues through the caesura fill. The dominant chord (prepared by viio7/V) strongly implies the tonic C minor, as does the caesura fill that follows it, but the 81 Charles Rosen, The Classical Style: Haydn, Mozart, Beethoven, expanded ed. (New York: W.W. Norton & Company, 1998), 208. 89 dominant is sustained through all of that until m. 82, corresponding to the moment in the exposition that signaled that the chance for an MC has passed. An ear trained in this style of tonal music would recognize that we are in C minor at this point based on the harmonies implied by the filler material, but C minor is not treated as an arrival point following from the dominant that precedes it—it just happens as part of the ongoing progression. Measures 86 until the ESC in 91 have more chordal textures that are clearly in C minor, paving the way for the PAC/ESC and the confirmation of the tonic key. This is perhaps the “lesson” of this piece: sometimes it’s not the moment that counts but the process. While there are a few clear cadences throughout the movement, they are often subsumed as part of larger spans. And there are several other arrival points, or approaches toward a possible arrival point, that do not provide the kind of punctuation that is the foundation of much of Sonata Theory. In this case, there is a disconnect between the strength of the clearest cadences of part 1 of the exposition (the PAC and HC of P) and the role they play in the overall form—they are not the most structural cadences of the movement. On the other hand, there are blurry, somehow compromised moments that seem to suggest a major formal event, but which dissolve or unfold gradually. The complexity of this movement permits different interpretations of the foggiest moments, but the first and second rotations are inextricably linked. 90 CHAPTER III THE DEVELOPMENT While Chapter II was primarily concerned with the middle of sonata expositions, this chapter will shift focus toward the outlines of large-scale formal sections. Essentially, this chapter explores the essentially binary nature of the galant sonata form, with part 1 consisting of the exposition and part 2 consisting of both the development and tonal resolution. It is the content and organization of the development specifically that opens the way for a Type 2 sonata in particular. Consider the following exposition, composed by Johann Stamitz c. 1752-1755, which comes from the first movement of the Symphony in D major, D-3. The V: HC medial caesura divides the exposition into two parts which are contrasted in terms of both key area and thematic content/character. There is, however, one element of this exposition that lends itself to complication: what are we to make of the opening four measures? The movement’s opening has implications not only for determining the structural downbeat of the movement as a whole but also for teasing apart its essential binary structure; those measures do not return again at any point in the movement and are most notably absent at the start of the development. Example 3.1 shows the opening of the exposition. 91 Example 3.1: Johann Stamitz, symphony in D major, D-3/i It is typical among Stamitz’s sonata forms to omit the expositional repeat. In the greater part of galant sonata forms, even without the repeat sign the division between exposition is readily identifiable by the return of opening material (and, in more unusual cases for the period, the repeat sign is present but not the opening material; this is not the norm). In the case of this movement of D-3, however, Stamitz omits both the repeat sign and the opening four bars at the start of the development, beginning that section instead with m. 5 in the dominant. We can recognize that the first four measures of the movement are, to some extent at least, parageneric—occurring outside of sonata-space proper. They constitute not so much a theme as a gesture with a simple, recognizable harmonic and rhythmic structure. We can label these measures as P0 to recognize their parageneric status and the theme beginning in m. 5 as P1. The development then begins in m. 53 with P1. This division of opening material into P0 and P1 can occur in any of the sonata types, 1, 2, or 3; this feature alone, like other expositional features, does not influence 92 sonata type directly. It can, however, have an effect on how we perceive larger structural divisions in the movement and that is highly relevant to our discussion of Type 2 sonatas. Specifically, we can consider how the P0-P1 construction affects the sense of two parallel rotations. Parallel openings to the expositional and developmental rotations are the norm in galant sonata forms. In such cases, both rotations begin with the exact same thematic material; the return to opening material following tonally and/or thematically contrasting S/C material is what signals the start of the new rotation. Without that thematic parallel, the sense of a new rotation based on the exposition’s referential layout is diminished. Stamitz’s omission of the customary expositional repeat, and the elision of C material into the start of the development, further mitigates the sense of rotation. In this chapter I consider the development’s role in articulating the birotational structure of Type 2 sonatas, both in terms of surface-level formal characteristics and deeper organizing principles—especially rotation. In the first sense, this essentially amounts to an examination of patterns of scope and sequence in galant development sections: what are the boundaries of the development section and how are they defined? Is there rhyme or reason to how the development unfolds, especially thematically? I also argue, in the spirit of Sonata Theory that galant developments are most completely understood not in opposition to exposition and tonal resolution, but rather as the beginning of a process—the beginning of the end of the entire movement.82 By 82 This is distinct from development-as-process as described by Charles Rosen as a “series of techniques of thematic transformation” that can also occur in the exposition and recapitulation (though this type of development also appears frequently in late-eighteenth-century and nineteenth-century sonatas. Rather, I 93 considering not just individual characteristics of the development in a modular way but also attending to how the development supports the rotational structure of post- expositional space, we can better see how the difference between Type 2 and Type 3 sonatas depends on much more than a single moment of tonal resolution. To support these claims, I also explore Hepokoski and Darcy’s principle of rotation and its applicability in the context of galant sonatas. The principle of rotation’s applicability to expositions and recapitulations is relatively straightforward—the prescribed order of thematic modules is generally explicit, especially in galant sonatas—it is less so where developments are concerned, as there is considerably more manipulation of expositional material and new material introduced. Nevertheless, understanding development sections in terms of rotational structure—allowing, however, for some flexibility in analysis—allows us to relate the contents of development to that of the rest of the sonata. Moreover, it demonstrates how a development is not a random series of events but rather an organized trajectory from beginning to end. In Type 2 sonatas, whose second rotation encompasses both development and tonal resolution together, this path of this trajectory takes on increased importance. My approach to development sections is informed in large part by Sonata Theory and by L. Poundie Burstein’s monograph Journeys through Galant Expositions. Burstein’s book draws on eighteenth-century writings to present an analytical lens filtering eighteenth- century form through the metaphor of a journey (rather than the more architectonic refer to the development’s participation in a movement’s large-scale rotational structure. For the former concept, see Charles Rosen, Sonata Forms (revised ed.) (New York: W.W. Norton & Company, 1988), 262. 94 approach that moves modules around like building blocks).83 As Burstein notes (and as Mark Evan Bonds noted before him), it is possible to integrate so-called “container” and journey metaphors into a single theory of musical form.84 In fact, Burstein cites Sonata Theory as “exemplary” in that regard.85 Burstein’s approach supports and enhances analytical descriptions drawn from Sonata Theory, though the approaches are by no means equivalent. For instance, both Burstein and Sonata Theory’s approaches are motivated significantly by motion toward cadences. Burstein’s Neo-Kochian approach conceives of Perioden that culminate in Sätze of varying strengths; for Sonata Theory, the cadential goals central to sonata form are the MC, EEC, and ESC. Burstein often invokes the metaphor of rotation and uses it to describe expositions, recapitulations, and developments.86 Unlike Sonata Theory, Burstein’s approach centers mid-eighteenth-century— galant—repertoire. His observations provide a basis from which we can find what is most useful and applicable from late-eighteenth-century-focused Sonata Theory and adopt or adapt its principles for earlier repertoire. While the repertoire, terminology, and 83 L. Poundie Burstein, Journeys through Galant Expositions, New York: Oxford University Press, 2020, 5–8. 84 Mark Evan Bonds, Wordless Rhetoric: Musical Form and the Metaphor of the Oration, Cambridge: Harvard University Press, 1991. 85 Burstein, Journeys, 7. 86 Burstein treats late-eighteenth-century developments more flexibly than Hepokoski and Darcy, both in terms of thematic ordering and modulations. His descriptions of galant developments are more detailed and hew closer to Sonata Theory’s Type 2 and Type 3 sonatas. See Burstein, Journeys, 225–228. 95 mechanics may vary between Burstein and Sonata Theory, however, their shared emphasis on trajectory and cadential punctuation complement one another well. Taking Burstein’s approach has a few benefits that are particularly important to this project. First, it treats the development as a complete unit—a single span from beginning to end, rather than a series of fragments, sequences, and modulations glued together. Another way to think of this is that it reinforces the idea of the development (and other parts of the sonata) as a trajectory, single span, or journey over the idea of the development as container, in which parts of the expositions are thrown into the same pot. By adopting the same kind of metaphor for exposition, development, and recapitulation alike, this structurally levels the playing field so that developments “get their due” as more than a hodge-podge of expositional materials. In turn, this contextualizes the normativity of the mid-century Type 2 sonata. Burstein’s observation that over the course of the second half of the eighteenth century there is a shift in importance from the thematic return (of the main theme) at the start of the development to that of the recapitulation, provides some necessary context for the Type 2 sonata as a normative option during the galant and the increased prevalence of Type 3 sonatas near the end of the century. With all of this in mind, this chapter explores not just the characteristics of development sections, insofar as they can be theorized, but also some of the large-scale issues with Sonata Theory that have led some to question the relevance of the Type 2 paradigm all together. While these issues appear to be related to the tonal resolution-versus-recapitulation question, we must in fact consider the foundation and organization of the development in order to address that matter. The principle of rotation as the underlying logic of the development is a fundamental element 96 of the Type 2 design. The “fate” of the Type 2 sonata’s explanatory power is inextricably linked with the rotational nature of development sections. This chapter is divided into three sections. In “Characteristics,” I examine the role of development sections in Type 2 sonatas and address the difference between a recapitulation and a tonal resolution. In “Rotation and the Type 2 Sonata,” I argue that defining the Type 2 sonata by its lack of a recapitulation does not adequately account for the role the entire second rotation (development and tonal resolution) plays in the sonata process. This results in a negative approach to the Type 2 sonata, one which ends up thinking of the Type 2 sonata with the same lens as a Type 3 sonata. A positive approach to the Type 2 sonata—what I advocate for—recognizes the form’s birotational nature, each rotation constituting a single thematic process. Rotation’s status a fundamental principle of Sonata Theory has drawn sometimes intense scrutiny from scholars, some especially wary of its application, or skeptical of its relevance to, development sections specifically. Perhaps the most thorough critique of rotations and Type 2 sonatas comes from Paul Wingfield’s 2008 review of Elements of Sonata Theory. I address these concerns with examples from eighteenth-century repertoire and make a case for its appropriateness—perhaps especially in the context of Type 2 sonatas. In doing so I offer support for the value of the rotational metaphor in sonata form analysis. Both of these sections engage both Sonata Theory and Burstein’s Neo-Kochian approach.87 Finally, I end the chapter with the analysis of movements from two Johann Stamitz symphonies, one in G major (G-2) and one in E-flat major (Eb-1). Stamitz’s 87 Burstein, Journeys, 10. 97 commitment to the Type 2 sonata provides an opportunity to explore some of the variety inherent in the form as well as the consistency of his preferences at different stages in his compositional career. In each analysis, I explore the formal idiosyncrasies characteristic of Stamitz and how they affect the larger form. Overall, this chapter makes the case for considering the development a fundamental part of the Type 2 sonata and sets the stage for examining the tonal resolution in Chapter IV. Characteristics The development section of a sonata form poses a unique theoretical challenge. Expositions lay out thematic material and a tonal trajectory established by convention; recapitulations return to that material with sometimes minimal modifications—often just enough to prevent modulation away from the tonic key. Developments, on the other hand, are almost by definition less predictable: they tend to modulate to other keys (aside from the dominant or relative major, depending on the mode of the movement), and themes and motives are manipulated—they may be fragmented or varied. Sometimes new material is introduced in the development that was not heard in the exposition. Conventions are much looser for developments than for expositions and recapitulations; even so, in the eighteenth century the content and events of the development are not random, and norms of thematic order and tonal areas do emerge upon study in the repertoire. “Development” can refer both to a formal section and to a process—a process that occurs in, but which is not confined to, development sections. The primary concern of 98 this chapter is with development-as-section and with that section’s role as the start of a rotation, though the process of development also influences the “Type 2-ness” of a movement.88 Analyzing a development section as rotational assumes that there is a normative layout with which it can be placed in dialogue—specifically, it assumes that this layout is, fundamentally, the same as that of the exposition. This is not to say that a development must, in Hepokoski and Darcy’s view, correspond point-for-point with the exposition that precedes it—and certainly, even a cursory examination of the repertoire would show that this is often not the case—rather, it places the development in dialogue with the syntax and overall trajectory of the exposition. This primarily concerns ordering: P comes before TR, which comes before S, which comes before C. Elements may be omitted, but the order in which those elements are presented cannot change. Most developments, including developments of this period, use more than just P material; the further we get through the referential layout (with thematic zones appearing in order), the more the rotational metaphor applies. In Type 2 sonatas, we can expect that the development will only cover part 1 of a two-part exposition; essentially, the point that we get to S in the referential layout is the latest point by which we must have reached the tonal resolution (an early crux with could send a clear signal about the arrival of the tonic key, but it is the tonal resolution and the dominant preparation that precedes it that confirms the tonic return). Some initial questions we can ask as we approach individual development sections are 1) how far does the development take us through the referential 88 See, for instance, Peter A. Hoyt, “The Concept of Développement in the Early Nineteenth Century,” 141–162 in Music Theory and the Age of Romanticism, ed. Ian Bent, Cambridge: Cambridge University Press, 1996. 99 layout? and following that 2) to what extent does it make sense to hear the development as rotational? Given the norms expressed in the repertoire, we can reasonably observe that the layout of most Type 2 developments a consisting of P/TR forms a half-rotation that is then completed by the second half (the tonal resolution)—S /C. (Hepokoski and Darcy note that P /TR developments are most common in sonatas generally, and that C is more common than S.89) Compare this formal structure with some of the options normally available to Type 3 sonatas. A complete development is an option, and it is common for developments to consist of half-rotations through part 1 (P + TR) of the referential layout.90 Additionally, Sonata Theory describes three other options: writing over (episodic opening), C-based opening, and S-based opening.91 These are all grounded in principles of rotation: writing over entails the substitution of new material over (usually) the first module of P—and then the rest of the rotation continues; C-based openings elide the end of the exposition with the start of the development, creating the effect of overlapping rotations; and S-based openings are rare due to S’s tendency to drive toward the EEC and ESC (i.e., its function of ending a process is somewhat at odds with the need to launch the development proper). 89 Hepokoski and Darcy note that P /TR developments are most common in sonatas generally, and that C is more common than S (Elements of Sonata Theory, 205). 90 Half rotations based on part 2 of the referential layout—S/C—are rarer and more problematic because they obscure the sense of rotation. See Hepokoski and Darcy, Elements of Sonata Theory, 217. 91 Hepokoski and Darcy, Elements of Sonata Theory, 212–217. 100 In my study of mid-eighteenth-century Type 2 sonatas, I have only found examples of writing over and episodic openings (for instance J.C. Bach’s W A6/i, analyzed in Chapter II); S- and C-based openings will be discussed below. Even episodic openings appear rarely in this repertoire (and, as in Bach W A6, such cases may emerge or exist in dialogue with other elements of the form analyzed in Chapter II. Hepokoski and Darcy identify episodic openings to developmental space in Type 2 sonatas as third- level defaults; they also note that while examples can be found of episodic openings in some of Mozart’s developments, they are not as common in galant works. My study of galant Type 2s supports this. Writing over “works” in a Type 2 concept because after the episode concludes, the trajectory of the rotation remains clear. We may emerge from the episode at a later module of P or in TR; either way, S is “properly” prepared. Theoretically this is possible with C-based openings as well, where C substitutes for an early module of P and the referential layout picks up where the episode leaves off. However, I have yet to find an example of a Type 2 sonata whose development begins with a C-based opening. Neither have I encountered any Type 2 sonatas with S-based openings of developmental space.92 The reason for this is a little clearer: for one thing, this would severely distort the balance of the sonata’s thematic content. The sonata as a whole would become P---S-----(EEC)---------------------------(ESC). This would be an extreme suppression of P (retaining its primary status only by virtue of its order). Such a sonata would not be based in rotation (any return to TR after the onset of S would have the effect of a retrogression. Even if C were included (as in an S + C half-rotation), this 92 S- or C-based openings can be present in the context of Type 3 to Type 2 conversion, a phenomenon discussed in more detail in chapters IV and V. Also see Hepokoski and Darcy, Elements of Sonata Theory, 376. 101 does not solve the fact that the thematic trajectory would still be retrogressive, or else the scale of the development section would have to greatly exceed the normal length for the style and time period. In sum, S-based developments are (often) non-rotational, and they “belong” to the Type 3 option, not the Type 2 option. When they do occur in sonatas we might call Type 2s, they are the result of Type 3 to Type 2 conversions. By and large, then, we can assume that any given development section in mid- eighteenth-century repertoire is most likely to begin with P-based material, or possibly with an episode. Moreover, P1 is most likely to initiate the development, rather than a later module of P—even if P1/P0 has an almost parageneric function (Hepokoski calls this the P-incipit-launch.”93 This 1) strengthens the thematic parallel between the first and second rotations and 2) emphasizes the launching of the second rotation as a new process (but one related to that of the first rotation). While a P-based opening is not, on its own, sufficient to establish whether a sonata will be a Type 2 or a Type 3, a developmental opening that is not P-based 1) significantly reduces the chance of a Type 2 sonata or 2) presents expressive complications that require further analysis (see conversion below).94 While the difference between binary form and Type 2 is often located at recapitulation, the development also has a part in this distinction—which also forms the basis for the distinction between Type 1 and Type 3) sonatas. Burstein notes that there is a shift in the relative importance of beginnings between galant and late-eighteenth-century 93 Hepokoski and Darcy, Elements of Sonata Theory, 373. 94 Hepokoski and Darcy, Elements of Sonata Theory, 373. 102 sonatas.95 This shift correlates to the waning popularity of the Type 2 form in the final decades of the eighteenth century. In this chapter, rotational developments are considered normative, though not as an inflexible or immutable axiom. For the sake of this repertoire, the rotational metaphor is particularly apt, and the musical events of these pieces lend themselves well to being heard this way. The question of rotational developments—and of the concept of rotational sonata forms more generally—in other repertoire, particularly that of the nineteenth century, remains a topic to be explored elsewhere. By understanding the galant development in rotational terms, we can better understand how the content of a specific piece of this period interacts/is overlaid with these components. Just as an expositional rotation has thematic, rhetorical, and tonal components, the same is true of the development. I will deal only briefly with the thematic parameter here, as its fundamental characteristics are fairly straightforward.96 Thematic The thematic organization of the development is governed by the exposition’s referential layout and the large-scale principle of rotation. In Type 2 sonatas, developments are based on P and TR if present, and the tonal resolution completes the rotation with S and C. There are two main differences between developments in Type 2 sonatas and those in 95 Burstein, Journeys, 227. 96 For more on development sections and Sonata Theory, see Hepokoski and Darcy, Elements of Sonata Theory, 195-230. 103 Type 3 sonatas: developments in Type 3 sonatas can consist of a full rotation through the entire referential layout, and they can also begin with S or C, a compositional strategy that is not only conceptually incompatible with the Type 2 model, but one which is not readily found in galant repertoire. Because movements’ second halves are often longer than the first, one might wonder whether an S-based tonal resolution in a Type 2 sonata necessitates a P-based opening of developmental space in the interest of balance or contrast; one could also think of it the other way around—that a P-based developmental opening necessitates an S-based tonal resolution. The first version of this line of thought is not compatible with the idea of rotation, and perhaps more importantly it treats a form of retrospective interpretation as an essential element of understanding a movement’s large-scale structure. This idea may be more profitably pursued in the context of nineteenth-century sonata forms. The second makes more sense in the context of this repertoire, though not for the same reasons: the progression from P to S in a Type 2 sonata is a matter of thematic trajectory baked into the second rotation. The implication that contrast or thematic over-saturation determines the content of either the development or the tonal resolution relies on ideas of the relative weight of one formal zone or thematic module over another; the fact that developments can heavily feature P or move on from it quickly suggests that there is no such calculation that can be systematically performed. The link between P-saturation and the Type 2 sonata specifically may come from an attempt to justify the perceived absence of P in the tonic following the development—there was so much P stuffed into the development that to bring it back yet again would be too 104 tiresome, the thinking might go. Such justification is unnecessary, particularly given the Type 2 sonata’s mid-century coexistence with an ample number of binary forms. While this may make sense for some movements, a more likely explanation comes from the nature of the expositional layout itself—especially in mid-century galant works. As I discussed in Chapter I, the scale of these movements is generally smaller than that of late-eighteenth-century counterparts; expositions are dominated by P and S and TR, if present, is generally brief and not necessarily clearly distinct. In order to project the parallel nature of the movement’s two rotations, P is likely to have a stronger presence than TR and the development—and S will be completely absent. New material is often introduced in development sections—often as a means of launching harmonic sequences and modulations. Especially in mid-century works, however, this new material is almost always introduced after at least the incipit of P has sounded, thereby establishing the parallelism between the two rotations. Ultimately, it has the effect of diminishing P’s power over the development while still preserving the essence of rotation. This view of development sections aligns more with the overall structure and history of Type 2s as a form—specifically from their binary nature. Retransitions may or may not use material drawn from the exposition. This affects the placement of the crux, the point at which the second rotation locks onto the exposition’s thematic track, which I will explore below. Rhetorical Thematically, the development section is normally governed by the referential layout of the expositional rotation. Both Hepokoski and Darcy and Caplin theorize that 105 developments are divided into stages based on rhetorical function as well. The rhetorical stages of the development are overlaid by the thematic layout of the exposition, and together these components provide aural landmarks for the listener. Hepokoski and Darcy and Caplin organize developments in different but complementary ways. Each identify action spaces whose functions differ from one another. I will take each in turn. In this section, I explore the different parts of an eighteenth-century development section using Sonata Theory and addressing Caplin’s theory of formal functions. Hepokoski and Darcy identify four zones that may be engaged in the developmental spaces of eighteenth-century sonata forms: the link, entry/preparation, central action zone, and the exit/retransition.97 The boundaries of these zones are determined not by the referential layout (i.e., no thematic module necessarily corresponds with any of the development zones listed above) but rather by the kinds of rhetorical devices employed: Link: optional, “precede[s] our sense of ‘the development proper’” (either zone 2 or zone 3 if zone 2 is absent) Entry/preparation: Also not mandatory. Often P-based, piano dynamic, variable length. Can be clearly demarcated or not. Central action: consists of 1 or more “events” or “parts.” Characterized by turmoil and shift to the minor mode. Commonly uses sequences through significant affective changes, but can also use intensified contrapuntal imitation, introduce a new theme or episode, etc. 97 Hepokoski and Darcy, Elements of Sonata Theory, 229–231. 106 Exit/retransition: preparation and execution of the dominant pedal. Each of these zones can have particular ramifications for Type 2 sonatas, though none is exclusive to or form-defining of this sonata type. I will discuss these ramifications later in this section. Like Hepokoski and Darcy, Caplin also divides developments into action zones: pre-core, core, and retransition.98 Building on Ratz’s theory, he conceives of developments as essentially built around model-sequence iterations which together form the “core” of the development. The core is normally preceded by pre-core material in which the development begins with material of lesser intensity. He notes that lengthy development sections often have two different cores: the first “confirms a development key” and the second “leads to the dominant of the home key to prepare for the recapitulation.” According to Caplin, the pre-core/core technique is very common in works by Mozart and Beethoven, but it appears much less often in works by Haydn.99 Hepokoski and Darcy identify their point of departure from Caplin’s pre- core/core technique as his reliance on the model-sequence formula as the primary organizing characteristic of development. They complain that his labels have no “larger explanatory purpose” and that it doesn’t say anything about the ordering of materials 98 Caplin, Classical Form, 141–159. For exceptions, see 155–157 (“Development Sections Without a Core”). 99 Caplin, Classical Form, 141 and 147. 107 within the development. In Classical Form, Caplin of course does not respond to Sonata Theory, written almost a decade later. Nevertheless, there is some overlap in their descriptions of developmental organization. Both Hepokoski and Darcy and Caplin talk about the prevalence of sequences in the development’s core or central action zone, something the listener can use to orient themself in time as a development progresses. The main take away about developments from both theories is that developments unfold in three general phases: 1) link and entry/preparation (Caplin’s pre-core), 2) central action zone (core), and 3) retransition. While both the Sonata Theory and form-functional descriptions of development sections are useful and relevant to the work at hand, I prefer Hepokoski and Darcy’s terminology for a few reasons and will adopt it for the remainder of the chapter. First, it provides a more specific framework for the events of the development. It is more explicitly tied to affect and affective changes in the development and less bound to sequence as a form-defining technique (while still acknowledging its relevance in this part of the form).100 Sequence is one possible means by which the tonal and thematic content of the development can be expressed, but it does not in and of itself affect rotational organization. Finally, the language of entry and exit Hepokoski and Darcy use is a particularly apt metaphor for the start of many galant developments, which generally ease into characteristic development procedures. Unlike the exposition and recapitulation, a 100 The sequence is still an important feature of eighteenth-century developments, even if as a technique it does not contribute directly to rotational structure. 108 development seems to me to be analogous to a freeway: Sonata Theory’s entry is like an on-ramp—it’s transitional to a certain extent, building in intensity but not yet requiring “freeway driving.” Once you reach the central action zone you’re on the freeway, employing a limited set of maneuvers to make your way forward. The exit is just that— you haven’t yet returned to surface streets, but you know that’s where you’re headed (if you want to take the metaphor further, the optional link is akin to an access road. In sum, Hepokoski and Darcy’s and Caplin’s concept of development sections is that there’s a section that is in some respect introductory (even though each author may have different criteria for what makes something “parageneric” in this sense),101 there’s a section that represents the “main” part of the development, and then there’s the retransition, which directs us toward the tonal resolution or recapitulation. Entry/preparation and retransition are the action zones with the most direct implications for the Type 2 form. As discussed above, the entry to the development of a Type 2 sonata is overwhelmingly P-based, reinforcing the birotational nature of the sonata. The retransition’s importance in a Type 2 sonata can be thematic and/or rhetorical, and it often contributes to a sense of tonal resolution over time as opposed to as a singular, discrete event. The retransition is especially important in Type 2 sonatas because it sets up the crux of the movement. Hepokoski and Darcy adapt this term from Ralph Kirkpatrick, who used it in his work on Scarlatti.102 According to Hepokoski and 101 Such links are uncommon in mid-century repertoire; most often there is a clean break between the end of the exposition and the P-based opening of the development. 102 Hepokoski and Darcy, Elements of Sonata Theory, 239–242, for the differences from Kirkpatrick’s approach, see 240 n. 11; Ralph Kirkpatrick, Domenico Scarlatti, (Princeton: Princeton University Press, 1953), 253–261. 109 Darcy’s usage, the crux is the point at which “the recapitulation can become, by and large, a simple matter of transposition.” In Type 1 and Type 3 sonatas, this usually happens at some point before the MC after material has been altered to prevent modulation to the dominant as in the exposition--it usually occurs at the end of P or with TR in Type 1 and Type 3 sonatas. From this point on, the sonata’s thematic material consists of expositional material at its original pitch level, so the moment from which the second-rotation tonic material corresponds to first-rotation dominant material is the tonic return of S. This is the case, for instance, in J.C. Bach’s keyboard sonatas (apart from a few more ambiguous examples). Given its prevalence in the repertoire, we can consider this the norm in mid-century works. Johann Stamitz’s symphonies represent a notable exception, however. A significant number of his symphony movements feature early cruxes—in other words, a crux that occurs before the return of S, and more importantly, before the caesura that marks the end of the development. Most commonly in Stamitz’s Type 2 movements, even a thematic correspondence between the exposition and retransition does not constitute the crux as the material is transposed to result in a I:HC compared to the expositional V:HC (the norm in Stamitz’s symphonies) In cases where the exposition was divided by a I:HC, the retransition often culminates in TR material at its expositional pitch level. In these instances, it is more appropriate to think of the tonal resolution as a process that unfolds over the course of several thematic and rhetorical events (the retransition, the caesura, and the launch of S). One of the effects of an early crux is to diminish the rhetorical importance of S’s tonic arrival. By that point, the tonic has already been 110 reestablished within TR. Does the tonic TR material (transposed from the exposition if preparing a V: HC) still constitute part of the development? Or, given that the tonic is not surrendered for the rest of the movement, does it constitute part of the tonal resolution? I argue that choosing either one of these labels misses the point—the two sections, distinguished from one another by their rhetorical function and role in the large-scale form, are elided. Hearing tonal resolution as a process that unfolds across that seam also emphasizes the birotational nature of the form as the most salient structural division is between exposition and development.103 For an especially notable example of tonal resolution as process in a Type 2 sonata, see the analysis of Mozart’s “Prague” symphony, K. 504, in Chapter IV. If we were to listen to a Stamitz symphony through late-eighteenth-century- centered ears, we would expect a recapitulation—a return of P in the tonic. In some Type 2 sonatas—namely, those with a clear caesura before the tonic launch of S—the division between development and tonal resolution is clean enough to seem like an appropriate place to “hang” a recapitulation. This view is not supported, however, by either Sonata Theory or Burstein’s historically-informed approach. Burstein notes that the point of recapitulation in galant sonatas—including the “true” recapitulations of Type 3 sonata— was often understated compared to the late-eighteenth-century recapitulation.104 103 As noted earlier in this chapter, Stamitz’s elision of the final expositional cadence into the start of the development is the exception, rather than the rule. 104 Burstein, Journeys, 222–223. 111 Tonal There are a few tonal aspects of eighteenth-century developments that have particular bearing on Type 2 sonatas. First is the expectation that the development will begin with P in the dominant and that the retransition will lead to a dominant pedal and, ultimately, the tonal resolution. While neither of these elements is by any means unique to the developments of Type 2 sonatas, these tonal landmarks are especially crucial when P does not return in the tonic for the remainder of sonata-space proper (and it strongly emphasizes the rotational nature of the movement’s second half and its parallel with the expositional rotation). Also of particular interest in the context of Type 2 sonatas are passages within the development that occur in the tonic before the retransition. When P returns in the tonic, a false-recapitulation effect may come into play.105 Not all thematic returns in the tonic within the development section necessarily evoke the false-recapitulation concept, especially when the thematic return is of a later module than P0 or P1. For instance, in the Stamitz symphony in D analyzed at the start of this chapter (D-3), the tonic statement of P material from m. 13 in the exposition in the second rotation (mm. 65–67) sounds less like a recapitulatory statement of P than part of a harmonic sequence. The global tonic is thus better heard as the subdominant of V, a flat-side step around the circle of fifths. There are a couple of factors that lead to this distinction: the fact that it continues, sequentially, from the material that immediately precedes it (mm. 61-65), and that it is situated within an often-changing tonal environment. This passage would not be mistaken 105 Hepokoski and Darcy, Elements of Sonata Theory, 221–226. 112 for a so-called “false-recapitulation,” even in the context of a Type 2 sonata with post-P- based tonal resolution. When considering a development section in a Type 2 sonata, we might first ask ourselves, does the capital-D-development begin right away after the end of the exposition or is there some kind of linking material? Consider, for instance, the development of J.C. Bach’s W A6/i. Note that this example is unusual (though certainly not deformational) in that its exposition is not repeated. What follows the EEC proper is not closing material, though neither is it particularly developmental (what I mean by this is that the four measures following the EEC (mm. 30-33) are a) extremely diatonic in III; b) divided cleanly into 2 + 2 measures with rhythmically parallel endings); and 3) an Alberti bass that clearly articulates a complete tonic-predominant-dominant-tonic progression). On the other hand, for all that they complete this progression, the four measures mm. 30-33 are also part of a longer passage that extends through a modulation to iv and PAC in iv in m. 39. In this case, the structure of the phrase supersedes the hypothetical autonomy of mm. 30-33, which are ultimately best heard as a subphrase within the larger phrase mm. 30-39. It functions somewhat like a “transitional introduction,” given the dual purposes of these measures.106 This development is its quite short and it feels like there’s not a single complete phrase of “real” development. As a section, it extends from mm. 30-47. It almost feels, however, as though there’s no real development proper: mm. 30-34 don’t feel primarily developmental, and mm. 40-47 feel 106 Caplin, Classical Form, 147. Note that Caplin uses this term differently, for cases in which the development begins in a key other than the subordinate key. 113 like retransition toward tonal resolution.107 In spite of its brief development, I would not call this a Type 1 sonata (sonata without development) because mm. 30-47 feel just developmental enough to avoid reducing that material to the status of a mere link between exposition and tonal resolution. This does, however, amplify the binary characteristics of the form (as opposed to its sonata form ones). Rotation and the Type 2 Sonata Analyses of sonata forms generally often focus on the exposition and recapitulation. This is not without reason, as it is the exposition that introduces us to the basic materials of a movement and sets the tone for what’s to come. Recapitulations bring us full-circle, as it were, and the return to the first theme—and especially to the tonic key—can be quite dramatic. Expositions and recapitulations are also relatively tightly organized and structurally predictable; they tend to follow clear patterns about which we can theorize systematically. Developments are another story. The processes and patterns of development are less tightly regulated by the conventions guiding the exposition and recapitulation, and the individual configuration and techniques of a development are harder to predict. We may expect, for instance, a theme to return in a foreign but closely-related key, but it is harder to predict precisely how motives may be placed in dialogue with one another, or exactly how they’ll be fragmented. 107 For a potential Caplinian solution, see “Development Sections Without a Core” (Classical Form, 155– 157). 114 Even so, there are patterns evident in mid-eighteenth-century development sections that we can follow, and even where such patterns aren’t evident, we can at least construct analytical frameworks that guide us through individual works. In Sonata Theory, development sections are treated as essentially rotational.108 In other words, they are fundamentally organized by the referential thematic layout just as expositions and recapitulations are. In Type 3 sonatas, the development constitutes its own rotation; thus, Type 3 sonatas have three rotations in all: one each for exposition, development, and recapitulation. In Type 2 sonatas, development and tonal resolution together form a single rotation. When scholars before Hepokoski and Darcy (and those who do not employ the metaphor of rotation in their analyses) analyzed Type 2 sonatas, they often identified them as binary forms of some kind. It is important to distinguish this from Sonata Theory’s birotational view of these pieces: while the A and B sections of a binary form correlate with the two rotations of a Type 2 sonata, conceiving of these pieces as rotational implies something about the organization within each of these sections as well. The underlying thematic ordering of the second rotation will align with that of the exposition’s referential layout. Under this model, the development and tonal resolution form a single formal unit on the same hierarchical level as the exposition. There are still three major processes that occur over the course of the Type 2 sonata form, however: exposition, development and return of part 2 of the referential layout. 108 Hepokoski and Darcy, Elements of Sonata Theory, 206–207. Regarding the possibility of non-rotational developments, see 207. 115 A significant number of works do have rotational developments, and straightforwardly rotational ones at that—especially in the galant style. The main signal is that these developments generally (more often than not) start with P material in the dominant and the rest of the material in the development comes from Part I of the exposition (P and TR)—with no material from S or C. Even when there is new material, it is usually “slotted in” without obscuring the overall arc of the referential layout from the exposition. The idea that developments in eighteenth-century sonatas are essentially rotational is more widely accepted than the idea of rotational developments (and, indeed, the appropriateness of the rotational metaphor in general) in nineteenth-century sonatas. Opposition to rotation has centered largely around nineteenth-century sonata forms, and they’ve brought the very existence of the Type 2 sonata (conceptually) as Hepokoski and Darcy see it into question. In 2008, Paul Wingfield provided an extensive review of Elements of Sonata Theory in Music Analysis. Wingfield was particularly critical of two concepts within Sonata Theory, which, not incidentally, are interconnected with one another: the concept of rotation, and that of the Type 2 sonata. In the following section, I respond to Wingfield’s main critiques and explore the benefits of rotational readings of Type 2 sonatas in pre-1780s repertoire. The metaphor of rotation itself is problematic for Wingfield, as the components of a sonata do not turn “in a plane through a given angle,” per Wingfield’s “scientific definition” of rotation.109 He would prefer, instead, the “more accurate scientific 109 Wingfield, “Beyond ‘Norms and Deformations,’” 145. 116 metaphor” of “periodicity.”110 But merely substituting one term for another does not address the full scope of Wingfield’s skepticism of Sonata Theory’s concept of rotation. He writes that “signing up to the rotational way of thinking is thus essentially an act of quasi-religious faith,” a statement implying lack of evidence and the expectation of unquestioning adherence.111 Finally, Wingfield takes issue with Hepokoski and Darcy’s treatment of developmental rotations in a structurally analogous way to expositional rotations.112 By responding to Wingfield’s concerns in turn, I can show how rotation and the Type 2 paradigm are not only connected, but also appropriate analytical lenses for mid-century repertoire. First, the matter of rotation becomes most controversial where developments are concerned. The thematic relationship between the exposition and recapitulation is explicit and, especially in the eighteenth-century, generally easy to follow. Both the exposition and recapitulation generally follow the same thematic path, and often the same general harmonic thrust without the modulation to a second key area. The relationship between the development and the exposition (or the development and the recapitulation) is not always as clear. There is generally more harmonic instability, more modulations, fragmentation of material from the exposition, and even entirely new material. How, 110 Wingfield, “Beyond ‘Norms and Deformations,’” 150. 111 Wingfield, “Beyond ‘Norms and Deformations,’” 149. 112 Wingfield, “Beyond ’Norms and Deformations,’” 150–151. 117 then, do such developments track onto the path of the rotational layout established in the exposition? In Sonata Theory, rotational developments are considered the norm. Hepokoski and Darcy introduce several concepts that account for common ways in which the sense of rotation is undermined, including writing over and S-/C-based openings of the developmental rotation. For Wingfield, the simpler option is to avoid the rotational metaphor entirely. His review seems to suggest an all-or-nothing approach to rotation and sonata form (developments specifically): either rotation is a central, immutable tenet of Sonata Theory, or it is not. The reality is a bit murkier. It is true that much of Sonata Theory hinges on the idea of rotation. I would stress, however, that what is important to Hepokoski and Darcy—and to me—is the form’s dialogue with rotation. In other words, we can conceive of a piece in rotational terms even when a literal or precise rotation is not in play. Sonata Theory does also allow for non-rotational developments, a more flexible approach than I think Wingfield sufficiently acknowledges. And even if their application of the rotational metaphor were overly strict, we as analysts could still derive benefit from adapting the principle for the analysis of eighteenth-century repertoire especially. It is beyond the scope of this project to determine whether the idea of rotation ought to be uniformly applied to all sonata forms, no matter the period or country of origin—though I highly suspect it shouldn’t be. I can say, however, that my own study of mid-eighteenth-century repertoire especially, the principle of rotation is highly salient, even in development sections. The strongest signal of a rotational development (though not sufficient on its own to determine whether a development is rotational or not) is a P- 118 based opening in the exposition’s secondary key (usually V). It means that we begin a new process within the sonata analogously to the exposition. The sense of starting something new is due not just to the exposition’s final cadence, or a change in dynamic, texture, or instrumentation, but by the sense of beginning something anew—material that would have been heard at least once, often twice, before. In the case of Type 2 sonatas, the remainder of the development normally uses only material from part 1 of the exposition: P and perhaps TR, if it is present. As discussed in Chapter II, this means that the development ends analogously to the MC point in the exposition, even if the material is not exactly the same. In the symphonies of Stamitz and Sammartini, and the keyboard sonatas and symphonies of J.C. Bach, one can expect to hear these modules in order. If new material is present, it still fits in to the overall context of the ongoing rotation (either as an insertion or, in Sonata Theory terms, by writing over some other thematic module from the exposition). Wingfield argues that the loose-knit nature of the phrases in development sections somehow precludes rotational organization. I argue that both can coexist within the development. Rotation is, after all, primarily about ordering—a sequence of events in dialogue with another such sequence. Whether themes are presented whole or in fragments does not destroy their relationship to the referential layout. It is true, however, that how clearly the underlying rotation can be perceived by the listener is important. And certainly there is great variety in the repertoire as to how overtly developmental rotations hew to the referential layout. 119 Conversions: It’s important to note, again, that formal types are not completely discrete and rigid. In the case of development sections, this is most often manifest in conversions between Type 2 and Type 3 sonatas.113 The idea behind conversion is that a process suggests a certain formal type but then becomes another. This is not necessarily the same as hybrid forms, though both hold some characteristics in common. Certainly in late- eighteenth-century works, and to a lesser degree in mid-century works, the Type 3 sonata is the most common of sonata types. Therefore, when one starts listening to a sonata, it is reasonable to expect it to unfold as a Type 3 sonata. This means that at some point, if we’re not in fact dealing with a Type 3 sonata, there has to be some point at which our perception of the form changes and recognizes a different formal type. In the case of Type 1 sonatas, this happens immediately (or soon after) the end of the exposition with the recapitulatory return of P in the tonic. In Type 2 sonatas, this conversion occurs later in the movement, often with the arrival of S at the tonal resolution. If the development opens with P, we do not have enough information to suggest we’re dealing with the Type 2 option; the Type 3 default stands. If, however, the development opens with S or C, that is an affirmative/active signal that the Type 3 option is engaged. If, following an S/C- based opening the tonal resolution still coincides with S, we are dealing with a Type 3 sonata that becomes a Type 2 sonata. The track of what begins as a Type 3 sonata is diverted onto that of a Type 2 sonata, which is where it finishes. The most important thing here is that both the Type 3 and Type 2 paradigms are engaged in such a 113 Hepokoski and Darcy, Elements of Sonata Theory, 376. 120 movement; the Type 3 characteristics still do not “belong” to Type 2s, and the conversion does not negate the characteristics of the first part of the movement. Johann Stamitz and the Type 2 sonata The Mannheim School of the mid-eighteenth century was renowned for its musicians’ technical skills and certain idiosyncratic musical characteristics. Johann Stamitz stands out as the central figure of the Mannheim School, not only for his involvement and leadership but also for the lasting impact of this music and musical culture on such esteemed figures as Wolfgang Amadeus Mozart. Stamitz’s symphonies, of which there are fifty-eight extant, plus ten orchestral trios, are composed in large part of Type 2 sonatas. The following analyses of the fourth movement of the symphony G-2 in G major and the first movement of the symphony Eb- 1 in E-flat major, illustrate how these movements’ two rotations unfold in the context of the idiosyncratic characteristics of Stamitz’s compositional style. Eugene K. Wolf’s study of Johann Stamitz’s symphonies remains an indispensable resource for getting to know the Mannheim composer’s style and works. While Wolf refers to them as “polythematic binary forms,” most of Stamitz’s fast symphony movements are Type 2 sonatas in the terms of this dissertation, while slow movements are most often simple binary forms.114 In addition to some of the characteristics frequently associated with the Mannheim style (“rocket” openings and the “[steam]roller” 114 Wolf, The Symphonies of Johann Stamitz, 139. 121 crescendo), there are a few characteristics that set Stamitz’s sonata forms apart from contemporary symphonists in other geographic regions. The repeat signs customary at the end of the exposition are frequently absent and the end of the exposition and start of the development are elided. In his later symphonies, Stamitz often reorders material, complicating the movement’s dialogue with rotation. Often in Stamitz’s symphonies there is a P0 module that sets up the key and the tempo. There is often a rhythmic component to this module, and they sometimes take the form of hammer blows. Analysis: Johann Stamitz, Symphony in G major, G-2/iv The third movement of Stamitz’s symphony in G major, G-2 follows two others that test the boundaries of formal type. In those movements, there are multiple approaches to the tonic that could be understood as the retransition, one of which introduces material from P and one of which introduces material from S (but with a heightened sense of expectation compared to its arrival in the exposition). By comparison, the third movement is much more readily identifiable as a Type 2 sonata. This movement is unique among the examples discussed in this dissertation for its use of a Baroque- and galant-style tonal layout that neatly clarifies the development’s thematic and tonal structure; as a result, the set-up of the tonal resolution is especially clear and overtly highlighted. This presto finale is relatively short and concise. P (mm. 1–17) consists of a single thematic idea whose last few measures are modified upon repetition to produce a PAC on V and lead into TR (mm. 17–38); TR ultimately leads to a V: HC MC in m. 39. The 122 effect of this is that very little time is spent in the tonic G major at the start of the movement: about twelve measures in 3/8 at presto simply flies by. Example 3.2 shows the Violin I part of the movement in its entirety. Example 3.2: Johann Stamitz, Symphony in G major, G-2/iv, mm. 1-41 The MC is clearly defined by first hammer blows on the new tonic and then hammer blows to articulate the V: HC MC. There is also an abrupt shift in dynamic from the forte of TR to the piano of S (which begins in m. 38). The contrast is further articulated by a change in rhythmic and melodic contour: while P is defined by an almost constant stream of eighth-note arpeggios propelling us forward both rhythmically and melodically, S begins on a quarter note as though struggling to keep the reins pulled in by the MC gap 123 in check. It is also characterized almost exclusively by stepwise motion, compared to the theme-defining arpeggios of P. There is one motive that is shared among TR, S, and C, which begins in m. 53: the turn figure that originally appears in m. 17 at the start of TR. This motive is especially prevalent in TR, less so in S (appearing twice in mm. 47 and 49), and regaining its prominence in C. It appears everywhere in the exposition except for P, and when the development begins in m. 69, it is noticeably absent—and it continues to be so until just before the tonal resolution. Figure 3.1 lays out the structure of the development. Of particular importance is the vi: PAC in m. 91. Such a cadence in the relative minor is a common device in Baroque and galant compositions, though often this cadence is followed immediately by the return of the tonic; in this case, the vi: PAC gives way not to the tonal resolution, but rather to the retransition (rTR). Figure 3.1: Johann Stamitz, Symphony in G major, G-2/iv. vi: PAC ends the development P------------(new)------- rTR------------- S V V/vi vi:PAC. V/I V:HC I m. 69 m. 77 m. 91 m. 92 m. 108 m. 109 This setup—that of a clearly articulated retransition, set off from the rest of the development—presages the tonal resolution both harmonically and, by implication, thematically (in another composer’s work, the imminent return to P material would be 124 implied; given the prevalence of Type 2 sonatas in Stamitz’s symphony movements, the odds are more balanced). Thematically, the development is dominated by P-material as well as the new musical idea that leads into the vi: PAC. The motive shared between TR, S, and C appears as well; most importantly, however, there is no other material that could be confused for S in the entire development. The melodic contour of the retransition (primarily descending stepwise) hews much closer to S; that it is separated by a clear caesura in this case from the rest of the development seems to approximate to some extent the divide between the two parts of the exposition. The emphasis on P in the development, coupled with the strong rhetorical gaps following the vi:HC and V:HC seem to lend some balance to the thematic content of the movement as a whole, and at the end of the movement, S has a chance to shine. Analysis: Johann Stamitz, Symphony in Eb major, Eb-1/i One of the characteristics that sets Stamitz’s Type 2 sonata forms apart from the majority of those of his contemporaries is that, especially in his later compositions, they are quite modular. In other words, thematic units from the exposition can be reordered in the development and tonal resolution. As a whole, this weakens the sense of rotation in these works and requires an adjustment to our expectations about the Type 2 sonata and the principle of rotation. Even so, the large-scale trajectories and outlines of rotation are still present in these compositions; in spite of significant differences between them and the 125 Type 2 sonatas of, say, J.C. Bach or Luigi Boccherini, many of the same guiding principles remain in effect. The first movement of the symphony in E-flat major, Eb-1, likely composed in 1754 or 1755, provides an excellent example of analytical issues specific to Stamitz’s symphonies. The formal structure of this movement is not as straightforward as that of D- 3 or G-2. Figure 3.2 shows the thematic components of the exposition: Figure 3.2: Stamitz Eb-3 Exposition Themes P0 P1a P1b P2 TR1 TR(P1a)2 MC-Fill S m. 1 5 11 18 31 39 46 51 I V/V V/V V/V V Init. Init. Cont Init. Cont Init I have divided P and TR into several units not because they contain several complete phrases each (as is the case in other analyses in this dissertation), but rather because A few things can be clarified straight away regarding the roles of certain materials Stamitz treats several motives within these themes modularly; labelling these motives within their thematic zones better reveals how rotation is undermined and yet still ultimately preserved. P0 is the most straightforward of all the thematic and motivic units of the entire movement. It serves exclusively as a framing device, opening the movement and closing it as well (with only cursory modification for a satisfactory end). P1 a is the first of the movement’s motives that returns again and again throughout. It has sentential qualities, a 126 two-measure basic idea revolving around the third Eb-G and descending second Eb-D (mm. 5–6) that is then chromatically sequenced (mm. 7-8), followed by an expanded continuation (mm. 9–15, with a tag—mm. 11–14—repeated in mm. 15–18). The continuation is almost a mirror image of the basic idea: where the basic idea presents first the third and then the second, the continuation first develops the second and then the third. The first time P2 (m. 18) occurs, it is as an insertion between two iterations of P2, the first one ending in m. 18 and the second beginning at the end of P2 in m. 23. It occurs again in m. 66 as the start of the development. While in the context of the entire exposition, P2 appears quite late, especially following two frequently recurring motives (P1a and P1b), it is nevertheless firmly situated within the primary theme zone. As such, it observes the letter of the law where rotation is concerned (i.e., the development still opens with P material). It is also true, however, that a return of P1a material would theoretically convey a stronger sense of rotation. Given that P1a and P1b occur multiple times over the course of the exposition, however—and the fact that it is harmonically already in motion with a series of secondary dominants, it is possible that this would have actually had the opposite effect—that the inner circularity of these motives would have instead obscured the large-scale trajectory of the movement rather than clarifying it. P2 is thus perhaps the best vehicle for conveying the sense of rotation in this movement, even as it would seem to undermine it. The story of this exposition can perhaps be best summed up as one of intrusion. P1a and P1b insert themselves into TR, and thematically this sends us in circles as listeners; as a result it leads us to reevaluate our place in the exposition’s trajectory at 127 every stage before we secure the dominant key area. This effect will be amplified in the second rotation with the addition of harmonic and thematic development. In this case, even the formal boundary between the end of the exposition and the start of the development is difficult to place. This blurring of boundaries results from the fact that there are many individual thematic modules that achieve almost motivic status— in other words, they occur all over the exposition, rather than being confined to a specific thematic zone. Often in eighteenth-century works, if thematic boundaries are unclear, one can at least default to finding large-scale boundaries like those between the exposition and development or the development and tonal resolution. On an even more conventional level, there is no repeat sign at the end of the exposition to unequivocally mark the boundary between sections. This is not is not at all unusual in Stamitz’s symphonies. What is less typical is the fact it is not immediately clear whether the material that follows S (the passage beginning m. 66) belongs to the exposition (in which case it would be C) or whether it launches the development. As noted above, this is in fact P2 launching the development. Measure 71 reintroduces the P motive from m. 15 in the dominant and tonic, a common developmental tactic that confirms that we are securely within developmental space. TR thus ends up being a brief link between the end of the exposition and a development that ultimately takes off with P. This is followed by the two-sixteenth-eighth note pattern in m. 78, which comes from m. 31. It is this motive that intensifies the tonal motion, feinting toward several different keys and leading to ascending and descending chromatic lines in m. 84. The chromatic ascent and descent are in essence an 128 interruption within the development’s TR. This passage then leads to the reprise of m. 112 in the tonic. As I read it, m. 112 marks the end of developmental space, but as I discuss in Chapter IV, the passage leading up to this point presages the return to the arrival of E-flat major. The dominant pedal is established in m. 92 and it uses material from TR, setting up the return to S (corresponding to m. 38 in the exposition. Note this motive also occurs as part of P (m. 5); I am treating it as part of TR because the material that comes after it is also from TR. A similar phenomenon occurs in Mozart’s “Prague” symphony, analyzed in Chapter IV). By the time S sounds in m. 112, our ears have already been reoriented toward E-flat as tonic. Add to this tonal foreshadowing the ambiguity of formal function of P1a and P1b and it becomes more difficult to pinpoint an exact moment of post-development material. When P1a and P1b occur, the first question is how those motives are functioning at that moment—do they correspond to their initial appearances or to their recurrences in the exposition? There is no question of developmental status through m. 89; in m. 90, the material from m. 27 (the very end of P2) appears in truncated form to avoid modulation. By this point, the tonal “work” of the movement is done; nothing else needs to be altered to preserve the key. The ESC occurs in m. 127 and launches TR yet again! The remainder of the symphony occupies coda-space, beginning similarly to the development (mm. 127–131 correspond directly to mm. 66–70). Instead of the later TR motive invoked in the development, however (m. 71–74), we jump straight to P, which becomes a complete 129 restatement of mm. 15–18. Then there is a coda to the coda, which brings back the opening gestures of mm. 1–4, which are absent throughout the rest of the movement. All of this results in a movement in which the outermost frame—the first and last measures of the piece—are very clear, but other seams are blurred by both tonal and thematic means. In some ways this movement strains against the tenets of Sonata Theory—with so many different motives shared between thematic modules, and with the elision of phrase endings and beginnings the rotational nature of the piece may not be immediately apparent. It is there nonetheless, however, and it is this rotational underpinning that supports the movement’s themes and motives—and nowhere more so than in the development. While it exists on a much smaller scale, this movement provides an entry point for my analysis of Mozart’s “Prague” symphony, K. 504, in the following chapter. 130 CHAPTER IV THE TONAL RESOLUTION In a sonata form, the recapitulation, in which primary theme material returns and the tonic key is restored, is a crucial part of the movement’s tonal and thematic trajectories. It is an event as well as a section, one that is often prepared, and from which point the thematic material is more or less familiar. While the recapitulation can hold this crucial role within a particular sonata’s narrative, it is not necessarily a form-defining characteristic. Consider that thematic return does not automatically constitute recapitulation in the classical sense; the recapitulation section of a sonata form is dependent on both content and context. One would not mistake a tonic return of P material in the first half of an exposition for the start of a recapitulation because of its context within the narrative trajectory of the sonata form as a whole. Even once we accept this there are several steps to pinning down just what it is that identifies a recapitulation. It’s not merely that it follows development; the Type 1 sonata has no development but it does have a complete recapitulation of expositional material, and the Type 2 sonata has a development but no recapitulation, as discussed Elements of Sonata Theory and outlined in this dissertation. Sonata Theory presents a positive way forward, defining recapitulation as complete rotational structure rather than by excluding thematic returns that do not meet the criteria. 131 One benefit of Sonata Theory’s rotational principle when applied to recapitulations is that it accounts for both the content and sequence of the thematic material in the final section of a sonata form. The bones of the exposition must be recapitulated post-development (in other words, not every thematic subunit needs to be recapitulated, but some part P and S must be heard in the tonic key after the development). This is a feature of many descriptions of sonata form, including Edward T. Cone’s “sonata principle.”115 What Sonata Theory adds to this is the element of thematic ordering as a consequence of rotation: a recapitulatory rotation must preserve not only the bones of the exposition’s content, but its order as well. P must come before S, in other words, and S can never precede P. Context thus takes on a role across the entire movement that is already granted to it in the exposition. As a descriptor of musical practice, this conception fits well—the vast majority of recapitulations bring back expositional materials in the order in which they were presented. When this is apparently not the case, there are several paths forward for the analyst, each of which carries its own theoretical, historical, and analytical considerations. Throughout this dissertation, I have referred to three such paths, i.e., formal paradigms that describe a movement in which tonal resolution does not coincide with the return of P-material. Broadly speaking, we can think of these as 1) binary form, 2) the Type 2 sonata, and 3) a modification or deformation of the Type 3 sonata, which involves either the omission of P as an intentional excision or the reversal of P and S material. As I have discussed earlier in this dissertation, the first two categories (binary form and the Type 2 sonata) are not mutually exclusive: a Type 2 sonata form is a type of binary form 115 Edward T. Cone, Musical Form and Musical Performance (New York: W.W. Norton & Co, 1968). 132 construed broadly. The latter two, however—the Type 2 and Type 3 sonata forms—are not compatible with one another insofar as they describe two fundamentally different structures (the first a birotational form and the second a trirotational form). Because the tonal resolution is the most immediately salient and obligatory moment of differentiation between the two forms, it is tempting to focus on this moment when distinguishing between a Type 2 and a Type 3 sonata—essentially distinguishing between a piece with a full recapitulation and one without. I have argued throughout this dissertation that the difference between Type 2 and 3 sonatas originates at a deeper, more large-scale level through the interaction of various pre-tonal-resolution events such as the medial caesura and the onset of the development, rather than on the single moment of whether the tonal resolution coincides with P or S material. I have also shown that our analytical preoccupation with the recapitulation as a form-defining event is perhaps a consequence of looking at galant repertoire through the lens of late-eighteenth-century repertoire, thus giving sonata form more generally a Type 3-oriented bias. In this chapter, I consider some complicated late-eighteenth-century pieces, how the boundaries between Types 2 and 3 are porous and what makes them so. I also address the idea of the “reversed recapitulation” and Type 2 is actually Type 3 issues in the context of 18th century rep, which should not be conflated with the 19th century versions of that discussion (which I will address in Chapter V). In a Type 3 sonata, the development ultimately gives way to the recapitulation. This third rotation in a sonata form fulfills two criteria: the return to and confirmation of the tonic key, and the return of some or all of the thematic material from the exposition in the same order in which it originally sounded (this is the so-called “double return,” where 133 the tonic is restored at the same time as the primary theme returns). The recapitulation is launched by primary theme material and proceeds along the track established by the referential layout in the exposition. There are exceptions, of course: for instance, the subdominant recapitulation in which P returns in the subdominant occurs with some regularity in mid- to late-eighteenth-century sonata forms (and in a number of Schubert’s sonata forms) delays the tonal return until after the restatement of P (at some point below I’ll address why I don’t treat Type 2s in the same way as these Type 3s—in other words, why the “tonal” difference is not the same as the “thematic” one). And it is not uncommon for phrases to be added or omitted (though usually not reordered) in the recapitulation. The fundamental characteristics of the recapitulation in a Type 3 (or Type 1) sonata express the principle of rotation by 1) moving through the referential thematic layout and 2) confirming the home key at the ESC (essential expositional closure, the first I:PAC of the rotation in post-P space). Something different happens in a Type 2 sonata. Essentially, the entire second half of a Type 2 sonata represents a single trajectory through the referential layout, beginning with the P-based opening of developmental space and continuing with S—by which point the main key has been restored. The point here is that Type 2 sonatas are not Type 3 sonatas that veer off thematically at the point of the tonal resolution. The difference in their paths begins earlier in the movement with the development, as discussed in Chapter III. It is essential, then, to see the Type 2 and 3 sonatas as having the same expositional backbone and diverging post-expositional trajectories. The tonal resolution is not really just a single moment in time, but rather the culmination (or near culmination, since the ESC is really the end of the process) of a process shaped by 134 rotation. This has expressive implications, and there is a vast range of realizations that affect the sense of return. In this chapter I make two main claims: 1) that a positive approach to the Type 2 sonata (in relation to the Type 3 sonata) is warranted, and 2) that, especially in galant repertoire, tonal resolution in Type 2 sonatas is characterized by a process rather than existing as an event in time. Of course on some level, everything that happens in any piece can be taken as a stage of its unfolding; I argue that the tonal resolution’s path in a Type 2 sonata is more explicit, as well as tied to the very nature of these pieces. In Sonata Theory terms, P is the “initiator of rotations” while S occurs within an ongoing rotation; since a recapitulation constitutes a complete rotation, it must begin with P.116 In a Type 2 sonata, then, the space after the tonal resolution constitutes the continuation of the rotation launched at the start of the development (which, as discussed in the previous chapter, normally begins with P). In other words, while a major tonal event occurs at this formal seam, a new thematic rotation is not launched at the same time. Even setting aside the idea of rotation, there is reason to avoid the concept of recapitulation in Type 2 sonatas. Caplin also questions the application of the recapitulation concept when the tonal return and P do not coincide (though he ultimately permits the use of the term because it is “so traditional”).117 While he does not employ the rotational metaphor, he does compare the (full) recapitulation of a sonata form with that 116 Hepokoski and Darcy, Elements of Sonata Theory, 65. 117 Caplin, Classical Form, 173. 135 of a small ternary form. For Caplin, the return of P is a crucial part of the distinction between small ternary and small binary; for Caplin, the distinction in sonata forms between P- and post-P thematic returns in the tonic is comparable to the distinction between ternary and binary forms.118 This chapter approaches Type 2 tonal resolutions with two perspectives in mind: first, that we should think of Type 2 sonatas in “positive” terms; that is to say, we should think about what Type 2 sonatas do rather than what they lack (a recapitulation). Second, that the tonal resolution is part of a process and should be treated as such rather than an isolated event. In other words, we can’t just look at what happens in the immediate vicinity of the tonal resolution, but rather observe how it relates to the rest of the sonata— what happens in the exposition and development, as well as what follows the tonal resolution. This chapter addresses various issues within these perspectives] and culminates in analyses of two movements that highlight their centrality to the Type 2 sonata: the first movement of Mozart’s Symphony no. 38 in D major, K. 504 (“Prague”) and the first movement of Marianna D’Auenbrugg’s Keyboard Sonata in E-flat major. In their own ways, each of these pieces bring together many of the different concepts discussed in this dissertation. The D’Auenbrugg sonata features a very straightforward exposition whose second rotation triggers a tonal resolution and thematic reprise that unfolds in stages. The “Prague” symphony has a complex expositional structure that has strong implications for the second rotation. It is situated in dialogue with both the Type 2 and Type 3 118 Caplin, Classical Form, 173. 136 paradigms, and how one reads movement’s relationship to those forms is dependent on first establishing a reading of the expositional structure. A Positive Approach to Tonal Resolutions Steven Vande Moortele, in his 2013 article “In Search of Romantic Form,” asks whether we ought to establish an entirely new set of expectations for nineteenth-century forms based on what we observe in the music, or measure what happens in the music against the norms of the eighteenth century. He frames the former as a “positive” approach to form—one that establishes the norms against which a work is analyzed on the basis of its own stylistic period—and the latter as a “negative” approach—using the norms of another period to analyze the work.119 This is anything but hypothetical: nineteenth- century sonata forms are frequently analyzed using the terminology, frameworks, and expectations of theories designed to explain late-eighteenth-century works—specifically those of Haydn, Mozart, and Beethoven. While Vande Moortele is engaging with this question as it relates to late-eighteenth- and nineteenth-century repertoires, this is also a question we must ask ourselves about the relationship between the late-eighteenth- century forms of Haydn, Mozart, and Beethoven and their early- and mid-eighteenth century predecessors. In other words, to what extent is it appropriate to apply Caplin and Hepokoski and Darcy’s theories to earlier composers, as well as those working outside Vienna between, say, 1780-1800? 119 Steven Vande Moortele, “In Search of Romantic Form,” Music Analysis 32 no. 3 (2013): 404–431. 137 This needn’t—and probably shouldn’t—be an either-or matter. Just as Vande Moortele seeks a “theoretical model in which the negative and positive approaches to nineteenth-century [in this case eighteenth-century] form can coexist,”120 so do I argue for blending positive and negative approaches to mid-eighteenth-century repertoire—and to the Type 2 sonata specifically. This dissertation is grounded in the principles of Sonata Theory, but only so far as its models are reflective of the music it describes. First-level defaults between the early works of J.C. Bach and the late works of Mozart are different, for instance, in the case of the coda’s thematic basis: Bach’s sonatas are much less likely to include a P-based coda, while Mozart’s are much more likely to do so. After analyzing the repertoire, I draw a clearer distinction between these practices than Hepokoski and Darcy do. Despite these differences, however, much of Sonata Theory does indeed work very well to describe what’s happening in these pieces. For example, the sonatas I have analyzed are overwhelmingly rotational, and the sense of trajectory is clearer than that of container [Burstein and Bonds]. That is to say, there isn’t always a clear distinction between P-space and TR-space, for instance, but the span from one cadence or formal marker is usually clear. This distinction between positive and negative approaches can be applied to any situation in which something is compared to an external norm. Indeed, the Type 2 sonata is often described in negative terms—that is to say, it is often described in terms of the Type 3 sonata, and defined by its lack of a recapitulation. This has a few negative (in the 120 Vande Moortele, “Romantic Form,” 411. 138 usual sense of the word) implications for how Type 2 sonatas have historically been addressed: first, it draws undue focus to the tonal resolution, and in fact this becomes the defining feature of the sonata. As I argued in Chapter III, this view risks obscuring the rotational process that drives the Type 2 sonata. Second, it also risks minimizing the Type 2 sonata, relegating it to a form used by the “lesser” composers of the mid-eighteenth- century before the genuine article is revealed by the “great” Haydn, Mozart, and Beethoven. To better address these obstacles, I advocate for a positive approach to the Type 2 sonata. Sonata Theory’s approach to the Type 2 sonata specifically is also largely a positive one—Hepokoski and Darcy stress repeatedly that the Type 2 sonata is not a lesser, incomplete, or deformational version of the Type 3 sonata. However, they also occasionally refer to the Type 2 sonata as the “sonata without recapitulation.” This is convenient as an identifier, especially to those who are unfamiliar with Sonata Theory. On a deeper level, however, this description undermines Hepokoski and Darcy’s own claim that the Type 2 sonata is its own thing—that it is not merely an alteration of the “normal” Type 3 option (note that this applies to the Type 1 sonata as well). By contrast, while I have argued elsewhere in this dissertation that binary-form analyses of Type 2 sonatas tend to ignore the ways in which these pieces are in dialogue with sonata form, they have the advantage of describing what these pieces are rather than what they are not. To apply Vande Moortele’s terminology here, a negative approach to the Type 2 sonata would view the Type 2 sonata as a deformation of the Type 3 sonata (i.e., a Type 3 sonata in which the recapitulation, or rather the recapitulation of P, is omitted). This is not such an issue with Type 2 vs. binary; for one, we don’t think of binary forms as some 139 kind of alteration of sonata form and for another, these sonatas still fit very much into a large-scale binary structure. Hepokoski and Darcy argue through most of their treatise for a positive approach to the Type 2 sonata—that they call it the “sonata without recapitulation,” which suggests a negative approach, is an anomaly and likely a kind of shorthand for convenience. In this project, I have shown that it is possible, and indeed, desirable to describe these pieces on their own terms. But it’s also not sufficient to define Type 2s as sonatas in which the tonal resolution coincides with the reprise of S. This shifts the focus such that it is this one event, the return of S in the tonic key, that defines the form of a Type 2 movement. My approach treats the resolution-reprise as one stop over the course of an entire rotation. What defines the Type 2 sonata, in other words, is the thematic and tonal trajectory of the second rotation (and the bi-rotational structure of the sonata), not just the precise moment of tonal resolution and thematic reprise. This is not a trivial distinction between definitions of the Type 2 form. Taking an “event-based” rather than a “process-based” reading has a profound effect on how we understand the larger trajectories of a sonata form—an effect I’ll explore in more depth in part 2 of this chapter. The following section addresses an interpretation of a certain kind of Type 2 sonata that has been especially contentious since the Elements’ release: that of the sonata with “reversed recapitulation.” This approach is essentially a negative one, as it defines the Type 2 sonata—through an event-based approach to the tonal resolution— through the organizing principles of the Type 3 sonata and treats it as essentially deformational. 140 The “Reversed Recapitulation” The basic idea behind the reversed recapitulation is that it is a “normal” (Type 3) or complete sonata form in which the recapitulation reverses the order of the primary and secondary themes. In other words, the main sections of the sonata (exposition, development, and recapitulation) exist fully—it is the order of lower-order modules that is altered (and this alteration does not override the principle of recapitulation). The exposition and development proceed as “normal,” and the usual retransition features (especially the dominant pedal and caesura) prepare the recapitulation like any other— the only difference comes after the caesura, when S enters instead of P. Part of the appeal of the reversed recapitulation model is that it accords more readily with the textbook conception of sonata form. The recapitulation is often considered one of the essential components of sonata form.7 Thus it is tempting to always associate the tonal resolution of a sonata (mandatory in eighteenth-century style) with recapitulation. And Hepokoski and Darcy’s “tonal resolution” is itself not completely satisfying as it does not seem to account for what is, indeed, a thematic return of an important module within the referential layout. Even so, the reversed recapitulation model is not compatible with rotation (certainly not as defined by Hepokoski and Darcy).121 If one were to try to use Sonata Theory terms to describe the reversed recapitulation for the sake of comparison, such an interpretation would fall under the category of the Type 3 sonata (with three rotations, 121 Caplin also urges caution with the concept of the reversed recapitulation, though for form-functional rather than rotational reasons. See Caplin, Classical Form, 173–174. 141 one each for exposition, development, and recapitulation). In fact, Wingfield’s review of Sonata Theory argues against the primacy of the rotational principle using pieces with “reversed recaps” as evidence (albeit with a strong focus on the relevance of Type 2s in the nineteenth century).122 The form of a Type 3 sonata with reversed recapitulation would be depicted as follows in Figure 4.1: Figure 4.1: Type 3 sonata with reversed recapitulation P---TR----S----C // P---TR----[S----C] // [TR]---S---C----P Alternatively, Figure 4.2 offers a reading the same piece as a Type 2 sonata with P-based coda: Figure 4.2: Type 2 sonata with P-based coda P---TR----S----C // P---TR----S----C // P Consider also the feature that introduces the possibility of a reversed recapitulation (as opposed to speaking of a binary form or incomplete recapitulation): the P-based coda. It is difficult to generalize about what these codas sound like. They range in scope from taking just the very opening module of P (as in Stamitz’s symphony in E-flat major) to something more extensive (consider the first movement of Mozart’s piano sonata in D major, K. 311—a complicated and problematic Type 2. Based on the thematic content of the coda alone, the longer the duration of P material in the coda, the stronger the case for reversed recapitulation. Hepokoski and Darcy identify the P-based coda as one of two 122 Wingfield, “Beyond ‘Norms and Deformations,’” 158. 142 main strategies for eighteenth-century codas, the other being a brief, emphatic closing module. According to Sonata Theory, codas are “parageneric spaces” that are separate from sonata-space proper.123 Type 2 sonatas may or may not have P-based codas; they are especially common in those of Stamitz and Mozart. P-based codas take on a special resonance in Type 2 sonatas (compared to Type 1 or Type 3 sonatas) because they present P in the tonic key for the first time since the start of the exposition. This has led some scholars to the reversed recapitulation reading. My intent is not to write off or entirely reject the concept of a reversed recapitulation. Further study, especially of nineteenth-century works, is needed to investigate the extent to which the rotational metaphor is appropriate in that later repertoire. I maintain, however, that the rotational metaphor does work well for the eighteenth-century repertoire explored here, and—on the flip side of that, the concept of reversed recapitulation is inherently deformational and the expressive connotations of such a deformation does not seem appropriate for galant works in particular. We needn’t close the book on the reversed recapitulation, but it is more meaningfully applied in a “positive” way. Some pieces Hepokoski and Darcy would analyze as Type 2 sonatas have been analyzed elsewhere as binary forms (consider any of Stamitz’s symphonies Eugene K. Wolf analyzes as “polythematic binary forms”) or not considered sonata forms at all. The distinction between binary form and the Type 2 sonata is not nearly as powerful as that between the Type 2 and reversed recapitulation, however, and the analytical 123 Hepokoski and Darcy, Elements of Sonata Theory, 281–283. 143 consequences of the latter far outweigh those of the former. After all, Type 2 sonatas are fundamentally binary forms; binary forms can be productively called Type 2 sonatas when they engage principles of sonata form like thematic differentiation, or recapitulation, or development (note that genre and time period may also influence our choices, as well as similarities and differences with other works by that composer). Important questions include, “do we learn something about this movement by placing it in dialogue with sonata form? Does it teach us something about the movement’s logic or mechanisms of expressivity?” The matter of the reversed recapitulation is more complicated and more urgent. The main conceptual issue is its incompatibility with the principle of rotation; on a smaller-scale but related issue, it raises questions about the role of P when it occurs after the reprise of S. It’s incompatible with rotation 1) because it takes elements out of order, 2) because P is the module that has the initiating function necessary to launch a new rotation, not S, and 3) because it undermines the function of the ESC. In the reversed recapitulation reading, the reprise/recapitulation of P in the tonic is part of sonata-space proper, i.e. part of the recapitulation and not part of the coda; the “work” of the sonata is not complete until after P has been recapitulated. By contrast, the Type 2 reading treats the reprise of P as part of the coda, not sonata-space proper. It follows the ESC, and C if present, and thus occurs after the work of the sonata is complete. And this view is not held by Hepokoski and Darcy alone; Caplin adopts a similar view of such P reprises.124 124 Caplin, Classical Form, 173–174. 144 Paul Wingfield, in his 2008 review of Elements, offered a strong critique of Sonata Theory and of the analytical value of a Type 2 sonata reading over that of a reversed recapitulation. While the review engages several aspects of Sonata Theory, it is particularly notable for its critique of the Type 2 concept, one which is tied up in a critique of rotation as well. I discussed aspects of this review in Chapter III, but I want to examine Wingfield’s attitude toward the reversed recapitulation concept more closely here. Most of Wingfield’s objections to the Type 2 concept focus on the nineteenth- century compositions Hepokoski and Darcy identify as Type 2 sonatas. I will address the nineteenth-century-specific portion of Wingfield’s argument in Chapter V. While his stance on this formal category in the eighteenth century is slightly milder, he nevertheless approaches it with some skepticism. He comments, for instance, on Hepokoski and Darcy’s “fundamentally ahistorical vision” of the Type 2 sonata as equally viable to the “all-conquering” Type 3 sonata, especially in Mozart’s oeuvre.125 Wingfield takes several of Hepokoski and Darcy’s examples and argues that they are more convincingly analyzed as either a) binary forms, or b) Type 3 sonatas with reversed recapitulation. His primary eighteenth-century example is Clementi’s Piano Sonata in G major, Op. 37, no. 2, published in 1798, and he is clearly more concerned about the Type 2 category’s applicability to nineteenth-century repertoire. Much of this hearkens back to his problem with the metaphor of rotation, which I addressed in the 125 Wingfield, “Beyond ‘Norms and Deformations,’” 156–157. 145 previous chapter; it makes clear that the concept of rotation is central to the concept of the Type 2 sonata. Thinking of P reprises after the ESC as part of sonata space is inconsistent with how such returns are interpreted in Type 3 sonatas. As Hepokoski and Darcy note, “it is not reasonable to claim that when such a tonic P-restoration occurs in a Type 3 sonata it is self-evidently a coda, while when it is found in a Type 2 sonata it is to be considered part of a presumed ‘reversed recapitulation.”126 Process and the Tonal Resolution In this section, I explore the various tonal-resolution-adjacent elements involved in Type 2 sonatas and how they arise out of the rotational processes in play throughout the sonata. Rather than treat the tonal resolution in isolation, this approach understands the tonal resolution as something that in inextricably linked to other elements of the sonata like the MC and the trajectory of the development. In the previous chapter, I explored how rotation is an appropriate and analytically meaningful organizing principle for (most) mid-eighteenth-century sonata forms. By associating the tonic return of S material (or post-P1 material) with “tonal resolution” rather than “recapitulation” I accomplish the following: 1) I show that “recapitulation” is defined in part by a thematic component as well as a tonal one (i.e. the return of P in the tonic key is what triggers the launch of the “true” recapitulation); and 2) by contrast, 126 Hepokoski and Darcy, Elements of Sonata Theory, 382. 146 “tonal resolution” is defined only by the return to tonic it heralds, without the suggestion that a new thematic rotation is beginning. It acknowledges a shift from development without attaching initiating thematic function to it. Thus, while the tonal resolution can have some elements of recapitulatory rhetoric, those elements alone are not enough to constitute recapitulation; indeed, some of these elements are present when we talk about MCs as well. In other words, context is also important. In this case, it’s the context of rotation that provides a framework in which elements that could participate in recapitulation take on their larger formal significance (i.e., recapitulatory function). It is important, then, for us to understand the elements of return that are not “purely” recapitulatory and can be applicable to tonal resolutions as well. The main thing that is unique to the recapitulation is tonal resolution coinciding with P. Any of the other elements we tend to associate with recapitulation—dominant lock to prepare for tonic return, shift away from developmental techniques and presenting themes in a more stable version—can also be applicable to the tonal resolution in Type 2 sonatas. We can divide the things that characterize recapitulations and tonal resolutions into three zones, as it were: preparatory (that’s the retransition), initiating (that’s the entrance of P in a type 3 or else the start of S in a two-part Anlage or post-P module in a continuous Anlage), and drive toward completion (toward the ESC). I’ll focus here on the latter two zones. Initiating zone: It’s often possible to identify exactly where S starts—in fact, it often coincides with the moment of tonal resolution. But it’s also often possible to detect at the 147 very least hints of the material preceding S (usually TR) before that moment. This occurs often in Stamitz’s symphonies and represents what Hepokoski and Darcy call an “early crux.” With what I call the “spontaneous S” option, the material immediately preceding the MC need not correspond with the material of the exposition. The difference between the effect of a “spontaneous S” or early crux is especially relevant in the case of Type 2 sonatas (over the Type 3 sonata) because there can be no early crux and the start of every recapitulation is in that way a clean start. The end of the development marks the end of a rotation, the recapitulation marks the beginning of one so the retransition is always marking the start of a new rotation. Such is not the case in Type 2 sonatas—the retransition cannot, by the definition of a Type 2 sonata, initiate a new rotation. And just as the degree to which the start of the recapitulation is articulated tell us something important about the role of thematic return in a given sonata, the same is true of tonal resolutions and reprises: it’s just that there’s a different set of norms and techniques available for Type 2 sonatas (like the early crux!). Drive toward completion (ESC): In the context of the Type 2 sonata, this matters especially as concerns the end of sonata space. We expect S (or TR/Fortspinnung in the case of a continuous exposition) to ultimately lead to the ESC and close sonata space. The overall rhetorical trajectory of the second rotation, then, is one that begins with P in the dominant and ends with the ESC in the tonic—at the deepest level that is all part of a single span. In mid-eighteenth-century works, the ESC is obligatory. There is no 148 requirement for an material beyond the ESC, though closing themes occur frequently. As far as codas are concerned, there are a few possible options: 1) No coda 2) “Tag”: not a full-blown coda, but something added to the very end such as two or three hammer blows 3) Coda based on any material other than P: could be new or gestural, could be an extension of C material. 4) P-based coda: return to the opening material of P, modified as necessary for the final cadence in the tonic at the very end of the piece. I’ve discussed already how the concept of the reversed recapitulation is incompatible with that of rotation. It also works against another principle of Sonata Theory, this one perhaps more readily salient: the ESC. The implication of the ESC is that the tonic is “secure” rather than “provisional,” i.e. it is confirmed by the PAC in that key. This is the major work of the sonata; while closing material may follow to reinforce it, its role is supportive rather than fulfilling something left undone. A true recapitulation of P following this cadence would not only flout the principle of rotation, it exists outside the sonata space whose ending is articulated by the ESC. Part of the discrepancy between the Sonata Theory approach and that of “recapitulation” advocates (in the context of Type 2 sonatas) stems from what I deem insufficient consideration of the role of content and context in defining a recapitulation. In other words, at issue is a definition of recapitulation that is based on specific 149 characteristics and the idea that the recap is the thing in the tonic that follows the development. Therefore, any musical event that checks those boxes is rightfully a recapitulation. The characteristics in question can generally be divided into three categories: 1) preparation (this would be the dominant pedal and possibly intensified fragmentation of themes and motives), 2) HC and caesura (marking the end of the development and the imminent arrival of…something), and 3) the (adjusted) restatement in the tonic of themes from the exposition. It’s not that these characteristics are analytically unimportant or irrelevant, but rather that they are not defining features of recapitulations exclusively, as I have discussed earlier in this chapter. And it is not necessarily desirable to stop at determining whether something is properly a recapitulation or a tonal resolution. Rather, there is much more to explore about the characteristics of these moments and how they fit into the narrative of the entire form. The above features are indeed characteristic of recapitulations, but they also need to be contextualized within the larger sonata process. We already apply them flexibly, sensitive to the varied ways composers treat thematic returns: we can recognize false and subdominant recapitulations for what they are, even though they don’t exactly “follow the rules” above. We are able to make those judgments by situating these returns within a context that is bigger than the moment itself (on the level of the second half of a two-part sonata form). Another consideration related to context has to do with the relationship with tonal resolution and the exposition—specifically to the analogous passage in the exposition around the MC. In both cases, S does initiate something: part 2 of the exposition or the tonal resolution. A Type 2 reading does not negate the fact that something new begins 150 with the arrival of S in either section. But observing that analogy has another consequence: just as the MC and arrival of S in the exposition do not signal the start of a new rotation, neither does the HC and arrival of S at the tonal resolution initiate a new rotation. The heart of all of this is that P is the key required to unlock a new rotation, absent that what occurs beyond the tonal resolution belongs to the same rotational process initiated by the development. Another reason conflation of tonal resolution and recapitulation is unsatisfying is that it begins from a premise of omission or deformation that is not, by default, warranted. This is the “A Type 2 sonata is a Type 3 sonata without a recapitulation of P” idea. This perceived lack of P can take two forms: 1) Type 2 sonatas are underdeveloped/incomplete forms that are not fully realized enough to have a “real” recapitulation, or 2) Type 2 sonatas are Type 3 sonatas where P is actively omitted for expressive effect. Aside from the ahistorical posture of these claims, they also make assumptions that are not universally sustainable. Type 2 sonatas are complete works, neither underdeveloped or forcibly manipulated by necessity. Both are value judgments that distract from the musical operations at work in individual compositions. Why should we think a movement with a complete recapitulation is necessarily better than one shaped by a single post-exposition span? This dissertation has shown how we can feel an organized, interconnected web of processes, events, and musical objects within these seemingly simpler structures by which ought to be recognized in their own right. Further, deformation as vehicle for expression privileges nineteenth-century music and imposes it 151 anachronistically as an ideal toward which this music is not even striving (i.e. Bach doesn’t have to be like Brahms to be beautiful). Analysis: Marianna D’Auenbrugg, Keyboard Sonata in E-Flat Major, i This sonata’s exposition is extremely clear; though multiple interpretations are possible for the nature of P (P or P0 + P1, depending on how much importance you want to grant the introductory quality of the opening measures. The distinction between P0 and P1 is significant in my analysis because of the differing textures and rhetorical signals sent by each of those modules). Example 4.1shows the opening these themes. Example 4.1: Marianna D’Auenbrugg, Keyboard Sonata in E-flat major, mm. 1-19 TR is extremely clear, is the arrival of the V: HC MC in m. 31. The harmonic rhythm intensifies, providing contrast with the I-V oscillation of P, and the musical action comes 152 to a complete stop at the arrival of the phrase’s final V chord. Example 4.2 shows TR and the MC. Given the clarity of the MC, it is clear where S (shown in Example 4.3) begins. However, there are two plausible places to hear the EEC (m. 45 or 54). In spite of this, the overall trajectory of the exposition is clear (and where there is room for interpretation, the various options are not wildly different). The second rotation of the movement, however, takes that clarity and dissolves it at a crucial point in the movement: in the phrases leading up to the tonal resolution. Example 4.2: Marianna D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 14-36 153 Example 4.3: Marianna D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 30-56 The movement is clearly in dialogue with the Type 2 form. P0 and P1 never return after they are stated in V at the start of the development. And the overall formal foundations of the movement as a sonata form are solid, i.e. it is never in doubt that the principles of sonata form (as defined by Hepokoski and Darcy) are in play. As with the “Prague” Symphony, however, the tonal and thematic rotations at work in the movement are not always moving together, resulting in the second rotation’s fuzzy middle. I would define this “fuzzy middle” as approximately mm. 67-76 (the core of which, and the fuzziest, is mm. 69-71). The material of the first part of the development is attributable to the opening of the exposition (P)—overtly in mm. 57-63 and more subtly in 64-67. Example 4.4 shows the beginning and “fuzzy middle” of the development. 154 Example 4.4: Marianna D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 57-76 Tonally, the main key area of the development is unquestionably the relative C minor. The global dominant Bb hangs on for only a few measures before chromatic bass movement in mm. 58-63 leads us to V in C minor. The passage from mm. 67-70 emerges from a deceptive cadence in C minor (an arrival on bVI) and drives toward the stronger half cadence that follows it. Via chromatically descending bass from the Ab of m. 67 through G supporting a cadential 6/4 (not unlike the chromatically descending bass movement of we arrive at a bass note Gb. Our ears likely hear this as viio7/V (in C minor); we would expect that harmony to resolve to V (creating a half cadence in the global submediant). However, the spelling of that chord—specifically the bass note spelled as Gb rather than F# presages something else. The Gb is pulled down to F-natural, and the G root is absent entirely. Instead, the end of this phrase, marked especially with a fermata, concludes on a viio6/4. This arrival is somewhat shocking, not only because of the unexpected, unorthodox progression, or the sheer instability of the terminal harmony, but also because motion comes to a rather sudden stop. In m. 69, sixteenth-notes become eighth-notes, but there is no equivalent slowing of the tempo through quarter and half- 155 notes—we move directly from eighths to a whole note. It’s a bit as though the rug is pulled out from under the listener. The pull from this this harmony toward resolution in C minor—specifically to a i6—is strong. After such a dramatic rhetorical moment, this is an opportunity for things to be “set right.” Instead, our expectations are circumvented yet again. The next two- and-a-half bars are firmly rooted in E-flat major, announced without ceremony. This “return” to E-flat is short-lived, however, devolving after two-and-a-half bars back to C minor. It’s a momentary glimpse of sunshine that makes the return to c minor—the revelation that we haven’t escaped its orbit—that much bleaker. TR rhetoric picks up in m. 76. This passage’s association with TR and its move away from C minor, make it a likely candidate for the start of the retransition (this makes it lengthy compared to the size of the development proper; another potential spot for the retransition would be m. 82. This placement is less ideal because it’s somewhat mid-stream, whereas other zones in the sonata coincide with phrase openings). Example 4.5 shows the retransition leading into the tonal resolution. The material of m. 82 until the end of the retransition is technically, new, although the constant stream of sixteenth-notes relates it back to the end of TR in the exposition (m. 27) that leads to the MC. In the retransition, however, rhythmic motion is slowed in m. 86 for three measures of standing on the dominant. The end of the retransition thus has the reins pulled back on it, while the end of the exposition TR plunges head-first into the MC. In the exposition, there are a full two measures sustaining the V chord before linking accompanimental material bridges the caesura gap into the onset of S in m. 31. 156 Example 4.5: Marianna D’Auenbrugg, Keyboard Sonata in E-flat major, i, mm. 57-76 In the retransition, a suspension delays the resolution to the V chord (m. 88); this resolution is then elided with the same caesura-fill heard in the exposition. It would be difficult to determine which moment is more dramatic: the arrival of S in the exposition or with the tonal resolution. There are several elements (tonal/harmonic, rhetorical, and thematic) that contribute to the drama of an MC/half cadence moment, and the strength of each in these two moments of d’Auenbrugg’s sonata vary. In the exposition, the energy gain is clear (half note + sixteenths becomes a constant stream of sixteenths) as is the sense of a break between the P + TR block and the linking material going into S. V/V is clearly insinuated, but it is punctuated by bass motion ^2- ^5-^1 alternating with I of the new key (in other words, that dominant function is operating here is clear, but V is embellished). Harmonically, in the retransition, the emphasis on V standing on the dominant is stronger—an uninterrupted pedal on ^5 from mm. 86-88 makes the articulation of the dominant more overt at this cadence than in the exposition. On the other hand, the delayed resolution in the right hand of the V chord so 157 that it coincides with the link to S mitigates the sense of coming to a complete halt and of bracketing off of the development’s material for a new, clean start with S. How these cadences fit into the hierarchy of those around them also affects the drama of these moments. The power of the tonal resolution in this movement is mitigated by the return to tonic earlier in the development (m. 71). That moment is striking, strange and memorable, even though the presence of the tonic is short-lived. Still, in context those brief moments in E-flat heighten the return of C minor—they are ultimately ineffective in producing the tonal resolution, whereas the retransition ending in m. 89 does usher in the tonic. Performance practice and the nature of the instruments of d’Auenbrugg’s day also influence the impact of these moments. The MC in the exposition comes at the conclusion of a virtuosic display of scales and arpeggios, and the dominant is attained on the downbeat of the measure and sustained. The upper voice enters the strongest tessitura of the eighteenth-century forte-piano or harpsichord during this passage. It is the link that follows that allows for a cool-down and diminuendo to start the piano S theme. The retransition leads into the reprise of S in quite a different way. Here it is the suspension that does most of the slowing. While the texture is not much lightened in m. 88, there would be some decay on the sustained downbeat, and the dynamic of the suspension’s resolution would be derived from that decay (the resolution itself on a metrically weak beat and it is shorter than the suspended chord before it, adding to the diminuendo effect). All this to say that in the exposition we’re sort of plopped down on to the resolution, sliding down the descending scales that precede it, whereas in the retransition, a diminuendo effect occurs before the cadential progression is completed. 158 This piece has a very clear exposition, characterized by a quasi-parageneric P theme and pervasive diatonicism. The development complicates this, first by introducing more chromaticism (even in the context of local c minor) and by avoiding cadences and using deceptive motion. That’s what the lack of leading tone resolution is—it’s a symbol of the reluctance to resolve of whatever metaphorical agent we want to assume is making the piece “go.” That Eb comes back briefly, and with S-related material, makes it seem like maybe we’ve succeeded in resolving, though we soon discover this is not the case. Eb is attained through a sequence, almost as though it sneaks in and the arrival of Eb is a non-event—even if the confirmation of it is given added emphasis. Analysis: Wolfgang Amadeus Mozart, Symphony in D major K. 504/i (“Prague”) In addition to clear-cut examples of either Type 2 or Type 3 sonatas, there are also cases of sonata forms that cannot be so clearly described as one type or another. These can fall into a couple of different categories: cases where the seam is unclear—it’s difficult to figure out exactly where the seam is, or perhaps P and S are related closely enough that it’s difficult to tell which theme returns at the tonal resolution; and cases in which an expectation is established for one type and then at some point we realize we’re dealing with a different type. I focus on the second category here. Hepokoski and Darcy address this phenomenon, which they call conversion, in their chapter on Type 2 sonatas.127 A conversion, in Sonata Theory terms, is a situation in 127 Hepokoski and Darcy, Elements of Sonata Theory, 376–378. 159 which a piece is going along giving signals of one type (for instance, an S- or C-based opening to the development, which is specifically associated with Type 3 sonatas) only for the expectations raised by those signals to be denied in some way. If a development begins with S or C, it is appropriate to expect that we are dealing with a Type 3 sonata. If, however, at the end of the development sets up a dominant arrival that gives way to something after P1, the Type 3 trajectory is abandoned for that of the Type 2 sonata. Hepokoski and Darcy cite an early Mozart symphony, K. 19, as an instance of Type 3=>Type 2 conversion; the first movement of J.C. Bach’s Keyboard Sonata in C minor, W A6, discussed in Chapter II, also uses a type of conversion as the tonal resolution begins by suggesting P (which would make this a Type 3 movement) but after only a couple of beats S takes over. While this occurs on a more local level around the development-tonal resolution seam (as opposed to K.19, which involves the entire second rotation), the shift in expectations in response to how the music unfolds represents a conversion from Type 3 to Type 2. In the eighteenth century, the overall binary structure of sonata form—whether Type 1, 2, or 3—means that the division between exposition (rotation 1) and development (or recapitulation, in the case of a Type 1 sonata) is usually clearly defined. More importantly for the current conversation is the fact that any blurring at this seam sorts itself out in time and does not usually cast long-lived doubt on the sonata type in play. At the seam between development and recapitulation, however, the repertoire sees more considerable ambiguity—a trend that continues and strengthens into the nineteenth century. This ambiguity is tied to rotation and can take several different forms: 160 1) P sounds in something that could constitute a thematic reprise, but which is compromised tonally or thematically (this does not include off-tonic recapitulations—especially, in the eighteenth century, in the subdominant—which are in dialogue with this kind of ambiguity but according to Sonata Theory principles are fundamentally Type 3 sonatas. In a Type 3 sonata, this iteration of P initiates a new rotation; in a Type 2 reading a scenario like this in which the development has been entirely P-based would indicate that the tonally or thematically compromised P is still part of the development). 2) The distinction between P and S (or P and C) is blurred. If thematic ideas from parts 1 and 2 of the exposition share similar melodic or motivic material, it may be difficult to tell if the development constitutes a full or half rotation. All of this can be exacerbated by cases where the exposition itself is difficult to parse. Hepokoski and Darcy suggest repeatedly that conversion is an immediate transformation that occurs “by simple fiat.” In their analytical examples, they use descriptors like “yanked,” “wrenching,” and “sudden,” how the music at the point of conversion is “impulsively ‘changing its mind’.”128 There is not always an element of violence or force involved with conversion, however. This is especially the case when the tonal resolution or recapitulation seems to roll out gradually. One well-known example of this is the first movement of Mozart’s Symphony No. 38 in D major, K. 504 (“Prague”). This movement is Hepokoski and Darcy’s lone 128 Hepokoski and Darcy, Elements of Sonata Theory, 376–378. “By simple fiat,” 376; “yanked,” “wrenching,” and “impulsively ‘changing its mind,’” 377; “sudden,” 378. 161 example of a Type 2=>Type 3 conversion (the inverse of the conversion discussed above).129 While I agree with the fundamentals of their analysis, I think there is much more to be said about the end of the development, retransition, and start of the tonal resolution. Most urgently, this involves the complexity of the expositional structure and its implications for the second rotation. In this analysis I show that the movement’s formal conversion is anything but sudden, and that even once the Type 3 paradigm is attained, it remains colored nonetheless by the Type 2 framework.25 This movement is also an excellent case study for some of the foundational issues explored in this entire project: it shows the benefits of thinking about form dialogically and how highlighting the characteristics of a single type, i.e., the Type 2-ness of the piece, throws its Type 3 characteristics into relief and vice versa. If one feels the need to choose a single formal label for this piece, “Type 3” would be more appropriate than “Type 2,” but better yet is to embrace the movement’s hybridity: that more adequately describes the formal processes at work. I show 1) that the Type 2=>Type 3 conversion interpretation not only describes the events of the piece, but also contextualizes the significance of individual moments (the return of tonic and the “real” recapitulation) and themes (within P and within S). This movement encourages us to ask which parameters seem to the primary motivators of the form, and what feels “normal” or out of the ordinary. Rather than ascribe tension to any “dissonant” formal elements or events, we might instead ask how the Type 2 and Type 3 paradigms are layered on top of one another. Finally, the 129 Hepokoski and Darcy, Elements of Sonata Theory, 378. 162 “Prague” symphony demonstrates that there is analytical value to the Type 2 sonata even where formal type is ambiguous. For the purposes of this analysis, we will set the slow introduction aside, but we mustn’t forget it entirely, as it too contributes to the grandness of the movement. We’ll begin with an overview of the exposition’s structure, then turn our attention to the development and recapitulation and how the second half of the movement emerges from the first half. Let’s begin by noting the sheer scale of this movement: it’s an extensive one by Mozartean standards. The scale of the movement is reflected not just by its absolute size, but also by the number of motives and mini-themes it contains Figure 4.3 provides a diagram of the movement’s form, adopting a Type 2=>Type 3 reading of the formal structure. Numbers with decimal points refer to modules of P—motives that return in other zones as well. TR1.1 in m. 71 uses material based on P1.1 in m. 37, S2.1 uses material based on P2.1 in m. 55, and so on. Figure 4.3: W.A. Mozart, Symphony in D m. 55 P2.1 Major, K. 504 “Prague” (i) m. 63 P2.2 m. 71 TR1.1 Measure ROTATION 1 m. 77 TR1.2 +1.4 Exposition m. 88 TR1.2frag m. 37 P1.1 m. 97 S1 + TR1.2frag m. 41 P1.2 m. 105 S1m (-TR1.2frag) m. 43 P1.3 m. 112 S2 + S1var m. 45 P1.1 m. 121 S2.1 m. 49 P1.2 m. 125 S2.2 m. 51 P1.4 m. 129 C1.1 163 m. 136 C2.2 m. 212 TR1.2 ROTATION 2 m. 214 P1.3 Development m. 216 TR1.1 m. 143 P1.3 m. 224 TR1.2 m. 156 P1.2 (+P1.3) m. 228 TR1.2 + TR1.4 m. 164 P2.1 (P1.2) m. 235 TR1.2frag m. 172 P2.1 (-P1.2) m. 244 S1+TR1.2frag m. 181 P2.2 m. 252 S1m (-TR1.2frag) m. 189 P1.1 m. 259 S2+S1var m. 192 P1.2 m. 270 S2.1 m. 195 P1.2 frag+ P1.3frag m. 278 S2.2 “Recapitulation” m. 282 C1.1 [ROTATION 3] m. 296 C2.2 m. 208 TR1.1 Example 4.6 shows the opening of the Allegro, including the first three modules of P. Rather than go into each individual module, of which there are so many, from the start, let’s begin instead with an overview of the entire movement. The exposition is divided into two parts, but the division itself is somewhat unusual: the MC of this movement is the V:PAC in m. 97, shown in Example 4.7 below. The material that begins in m. 97 is a sharp contrast to the material of P in motivic content and texture, and it is firmly-rooted in the dominant. This is not to say that there are not prominent half cadences earlier in the movement. In fact, the I:HC in m. 71 demonstrates several characteristics of an MC, from the immediate textural break and the hammer blows in all voices in m. 70. The thematic material that follows, however, quickly undermines these qualities. Rather than proceed to new material, the opening material of P1 resumes. 164 Example 4.6: W.A. Mozart, “Prague” Symphony (i), mm. 37–43, Modules 1–3 of P As Hepokoski and Darcy note in Elements, this reveals that part 1 of the exposition is a large-scale period divided into “grand antecedent” and “grand consequent.” P constitutes the grand antecedent and, following the half cadence of m. 71, TR—which is based on P material—constitutes the grand consequent. The latter half of the exposition is punctuated by several V:PACs. In m. 71 we have our first significant cadence of the movement: a I:HC. The second significant cadence is the V:PAC that arrives in m. 97. It may first appear that these cadences represent the MC and the EEC; further inspection, however, offers a different interpretation. Listening in time, one could be forgiven for—even expected to identify— this cadence as the medial caesura, a second-level default in a movement of this scale. 165 Example 4.7: W.A. Mozart, “Prague” Symphony (i), mm. 91-100, MC and start of S The MC effect of this cadence is undone by the thematic material that follows: P material resumes, though in V, before characteristically-TR material begins. This is a clearer version of the grand antecedent effect operating in J.C. Bach’s Keyboard Sonata in C minor, W A8a (iii) analyzed in Chapter II. Once we hear TR-like material, we expect a HC MC to follow (whether for the first or second time, as in a 3-part exposition). In this case, however, we’re met with a V:PAC. Such a cadence could act as either a lower-level default MC option or as the EEC (as in a continuous exposition, subtype 2 as discussed in Chapter II). The latter can be ruled out after listening to the material that follows, which has a phrase structure and function far beyond that of merely reiterating the PAC as that interpretation would require. Most importantly, we don’t know what this cadence is by hearing it alone—it is 166 only as the next few measures progress that it is clear we’ve started something entirely new. Usually it is clear that the MC is the MC even before S begins—we can’t know that for sure, and certainly there are exceptions (consider even the first HC in the Prague symphony (i)—it has several MC hallmarks, and it is the P-based material that comes after it that reveals that we are still in P. This is even more ambiguous in that the cadence would be a lower-level default as an MC. In some ways it’s like a microcosm of the Type 2-Type 3 conversion that occurs later in the movement. If I were to identify an overriding theme for the narrative of this movement, it would be one of layers, of opening and closing doors in some kind of dance between thematic content and role and cadential closure. To summarize the structure of the exposition, it is overall in two parts. Part 1 unfolds as a large-scale period, with P as grand antecedent and TR as the P-based consequent. Part 1 ends with a V:PAC, a lower-level default, and C incorporates P material as well. One might describe this as a P-heavy exposition given both its extensive use of motives from P but also the role of P’s events on the structure of the rest of the sonata. Analysis of the remainder of the sonata form reveals why it is critical to think of and treat Type 2 identity as a matter of large-scale processes; in other words, it’s not just about S and it’s not just about the tonal resolution, but about how these elements interact with the movement as a whole. The opening of the development uses a gesture from P (I’m calling it P1.3, which corresponds to the high winds in m. 43 in the exposition). Technically this still reflects the norm of a P-based opening to the development, though an opening that would convey an 167 even stronger sense of rotation would have opened with P1.1. The rotational aspect is preserved insofar as the development begins with P, though this sense is compromised by the fact that it does not open with the very opening gesture of m. 37, thus lacking the P zone’s incipit. And in fact, the very opening module of P is studiously avoided throughout the development. Example 4.8 shows the opening of the development as Rotation 2 begins and as motives P1.2, P1.3, and P2.1 enter. Example 4.8: W.A. Mozart, “Prague” Symphony (i), mm. 140-162 168 The development ultimately employs each of the modules from P. In m. 162 we move into P2.1, and it is within P1.2 that the “search” for the tonic begins (we ultimately 169 reenter tonic space in m. 187, though D major is not yet secured by means of a PAC). In 189 we reach the half cadence that separates P and TR in the exposition—the half cadence separating P as antecedent at TR as consequent. The sense of recapitulation is undone by, first, our understanding of that half cadence as it occurred in the exposition (i.e., that it is not the MC); second, by the consequent’s move to the tonic minor; and third, by the fact that a dominant pedal undergirds the whole passage. This creates a situation in which we have a point of arrival at the half cadence, but, just as in the exposition, unfinished business remains. In this passage beginning in m. 189, this unfinished business is as much tonal as it is thematic. Measure 189 reintroduces P1.1 and P.1.2 follows shortly thereafter in m. 192. By both following the exposition part 1’s HC and the small alterations to bring it into the minor mode (comparable to the chromatic alterations of the TR consequent version in the exposition), it makes sense to hear this passage as TR. In this context, TR material serves as retransition out of the development and into the tonal resolution proper. All of this is complicated by the fact that from about m. 177 on, the ear gravitates toward D as tonic. Indeed, the repeated V-I gestures and tonic arpeggiations of P2.5 (mm. 181 and following) and the HC of m. 189 mean that by the time we get through the dominant pedal to the thematic return in m. 208, D’s centrality is firmly established in our ears. The change to the minor mode introduces an element of uncertainty, but it doesn’t require any modulation pyrotechnics to restore the tonic major. Still, the sense of tonal arrival in m. 208 is powerful, given that it comes out of such buildup of anticipation, while the initial restoration of the tonic in m. 177 appears with much less ceremony, almost as if arriving by stumbling through. It makes for a more satisfying sense of 170 recapitulation in m. 208 than in m. 177. The result is the sense of a tonal resolution that unfolds in stages, “completely” resolving in m. 208 but with one foot through the door thirty measures earlier. The matter of thematic return is even more complicated. The sense of tonal repose is so strong in m. 208 that when the first violins begin their syncopation and the lower strings take up the melody, it isn’t difficult to hear this as a normal recapitulation of P in the tonic. It’s a matter of what musical material is employed, and of our generic, stylistic, and historical expectations for this moment in a sonata form. Indeed, several published analyses place the recapitulation in m. 208 without much comment, including Hepokoski and Darcy.130 For the latter, this marks the point of conversion, the point at which Mozart could have introduced S for an unequivocal Type 2 layout but instead wrote a recapitulation. There are a few elements that undermine the Type 3-ness of this movement, however. They stem from the fact that really it isn’t P that returns in m. 208, but TR, as though TR gets a second chance after its minor-mode version in m. 189. I identify this passage as TR and not P for two reasons: first, because of the chromatic alterations in the melody that correspond to those of TR in the exposition (the A#s in mm. 209-211), and second, because the thematic process it begins leads not to the HC of the exposition’s part 1, but rather to the PAC MC (in this case a I: PAC instead of the exposition’s V: PAC. In a different piece, the evidence above would present a compelling case for a strict Type 2 reading of the movement. What makes the situation more unstable in this 130 Hepokoski and Darcy, Elements of Sonata Theory, 378. 171 movement is the fact that P and TR use such similar materials, and that in the final D major section it is difficult to determine which is in fact in play. There are still, for instance, elements of P not associated with TR in the exposition that crop up after m. 208 point, namely P1.3 in m. 214, and chromatic inflections give the air of a last-minute conflict before really settling into D major. To force a Type 2 label on the movement without further exploration of its dialogue with the Type 3 sonata is to diminish the impact of the movement’s second half tonally, thematically, and rhetorically. In cases where the tonal resolution begins with material that contrasts starkly with P, the tonal and rhetorical strength of the arrival should not be lightly overridden. What makes this movement a special case is the relationship between P and TR—both of which form, through their antecedent-consequent relationship, a tight unit as part 1 of the exposition. At the very least, this is a Type 3 sonata whose Type 3-ness is compromised. There are a couple of ways to think about how this happens. In this case, the latter part of the second rotation seems to have trouble getting back to the tonal resolution, which is revealed in fits and starts. While it is not uncommon for there to be a recurrence of the tonic within developmental space, the timing, number of “false starts” and the thematic content of those false starts gives this the sense of there being some kind of obstacle, a hill that needs to be cleared before the tonal resolution is truly, permanently achieved. In terms of timing, we have two false starts that occur after the development has had ample time to run its course. In other words, proportionally speaking, it is conceivable that each of them could represent the tonal resolution. That this happens twice is also notable: here it comes! Oh, no, wait, now here it comes! Oh, no….In terms of content, the module used in each of these moments of D major restoration comes not from the start of P but from 172 the middle of it. It’s also a module that appears at the end of S. That so much material is shared between P and S blurs the line between them—especially since the exposition’s MC treatment is not the most typical of the period. The uniqueness and breadth of S (that is consists of complete phrases and can stand on its own) is what most clarifies its role within the exposition (note that this is still in some ways more related to the cadences that surround and constitute S than, say, characteristics of its melody. To call the first movement of the “Prague” Symphony a Type 3 sonata without any explanation or qualification misses some of what makes the second half of the piece so exciting. Speaking of conversion is a step in the right direction. I prefer to speak of a hybrid Type 2-Type 3 form that is generated through gradual conversion/conversion in stages, as conversion suggests that the movement’s Type 2-ness is cast aside in favor of the Type 3 sonata. Rather, I argue that in the “Prague” symphony, both of these paradigms continue to operate through the tonal resolution. The salience of the P theme and that it is in fact based on P at the “recapitulation” point (m. 208) means that the Type 3 label is probably the best one. In a way, this is almost like a label of convenience more than representing what is actually going on, i.e. obscuring the motivic relationships that ultimately suggest Type 2 organization. But we needn’t—indeed shouldn’t—stop our analysis after this decision. This movement’s complex dialogue with the Type 2 form is essential to its construction (and a reminder that the paradigm isn’t irrelevant, even in 1786 when the symphony was composed). And note that none of this would be an issue if not for the structure of P and the recycling of P material past the P zone. Another thing that allows all of this to work is 173 how contrapuntally compatible the motives and themes of the movement are. It essentially allows the timeline of the movement to collapse on itself. The rest of the exposition proceeds in a fairly straightforward manner, with the notable exception of the return of P material at the end of S and as the basis of C. Even this doesn’t feel especially because it is quite contained: once P regains its grip it maintains it through the end of the exposition. This analysis focuses on Part 1 of the exposition, because it is the structure of the grand-antecedent and consequent around which P and TR are based that complicates the movement’s dialogue with the Type 2 and Type 3 forms later in the movement. One of the things that makes the tonal resolution here so special is that the version of TR it begins with is so closely related to P it feels a bit overly precise to not recognize that thematic entrance as a “real” recapitulation. Note that this distinction between Type 2 and Type 3 in this case is, as noted early in the chapter, about both content and context. The tonal resolution is also explicitly processual: at the end of the development D major is established before an excursion to D minor seems to interrupt the confirmation of the home mode. The recapitulation proper almost seems to start in medias res, but the pedal point under the minor version of P rejects recapitulation and reopens retransitional space. It’s only after all that that recapitulatory function is secured. we experience tonal resolution unfolding in time as opposed as a single moment that is achieved and immediately absorbed into our hearing. I argue that the “Prague” Symphony is technically a Type 2 sonata but that this is a situation where our experience in time seems to much more convincingly pull toward a Type 3 reading. On the other hand, one could argue that this is, in and of itself, a result 174 of Type 3 bias; perhaps if we didn’t consider Type 2 to be a lower level default the movement’s dialogue with the Type 2 paradigm—especially around the tonal resolution—might be more readily apparent. 175 CHAPTER V CONCLUSION In this I have explored how “Type 2-ness" is constructed from sonata processes, rather than being defined by a static formal landmark—the tonal resolution coinciding with secondary-theme material. I have also argued that the Type 2 sonata opens up pathways to explore the intersections and points of departure between the norms and principles of galant sonata forms and those of the late eighteenth century. As such, we should not so easily ignore the Type 2 sonata and its inner workings (or, worse still, dismiss it as an underdeveloped precursor to the “real” Type 3 sonata). Fortunately, more and more scholarly work centers galant repertoire and treats it as a style worthy of study for its own sake, rather than merely for its position between the late Baroque and High Classical. The Type 2 sonata is thus finally beginning to get its due in the study of eighteenth-century form. Beginning in the 1770s, the prevalence of the Type 2 sonata was greatly diminished. Hepokoski and Darcy hypothesize that this might have something to do with changing stylistic expectations, specifically a heightened expectation for the kind of formal drama that is often present around moments of recapitulation. Similarly, L. Poundie Burstein describes a shift from the galant to the late-eighteenth century from stressing the start of the development to placing more rhetorical/dramatic emphasis on the start of the recapitulation. While Hepokoski and Darcy provide examples of the Type 2 sonata in the nineteenth century, the form’s reduced prominence has generated significant skepticism 176 of its validity as a formal designation in this later period. I discussed some of the counterarguments to Hepokoski and Darcy’s claim that the Type 2 sonata is an appropriate framework for nineteenth-century sonata forms in Chapter IV, but it is worth taking a closer look at this matter and its potential implications for the new Formenlehre more broadly. As discussed in Chapter IV, scholars like Paul Wingfield and Steven Vande Moortele, have rejected the Type 2 sonata as a lens through which to hear nineteenth- century sonata forms. In most cases, these scholars prefer to hear such movements as sonatas with “reversed recapitulations.” This takes an essentially negative approach to these works, which are conceived of as Type 3 sonatas that have been subjected to a particular transformation: the reversal of the primary and secondary themes in the recapitulation. The Type 2 sonata thus represents a microcosm of the conversation that is currently underway amongst Formenlehre scholars: to what extent is eighteenth-century Formenlehre a useful theoretical and analytical reference point for nineteenth-century music? Conversely, to what extent is it necessary to develop a new Formenlehre for nineteenth-century music? As Steven Vande Moortele argues in “In Search of Romantic Form,” it is likely that the most meaningful path forward consists of a bit of both. Franz Schubert’s sonata forms offer particularly good testing ground for future exploration of how eighteenth- and nineteenth-century Formenlehren might interact. There are two related sonata form practices in Schubert’s works that are of particular relevance to an examination of the Type 2 sonata’s utility in the analysis of early nineteenth-century sonata forms: 1) off-tonic recapitulations and 2) thematically- 177 and rhetorically- ambiguous recapitulations. Both characteristics avoid the double-return of tonic and primary theme material. As far as off-tonic recapitulations are concerned, Sonata Theory considers them to be “true” recapitulations, and thus Type 3 sonatas as long as P still launches a new rotation. The subdominant recapitulation—by far the most common type of off-tonic recapitulation—has roots in eighteenth-century music; perhaps the best-known example is the first movement of Mozart’s piano sonata in C major, K. 545. While Hepokoski and Darcy consider sonatas with off-tonic recapitulations to be Type 3 sonatas, their delaying of the tonal resolution conceptually weakens the power of the recapitulation. While they are often like Type 3 sonatas in other ways (i.e., using typical retransition characteristics, clear dominant preparation—albeit in the dominant), they are more vulnerable, as it were, to a compromised sense of recapitulation. In other words, if there is not a clear caesura/rhetorical break between the end of the development and the recapitulation (see thematic ambiguity below), it may be especially difficult to pinpoint where the development ends and where the recapitulation begins. Thematically ambiguous recapitulations, especially in Schubert’s hands, can be more analytically challenging. In these sonatas, the return of P material is compromised in some way, thus blurring the line between development and recapitulation is blurred. This ambiguity can take several forms, such as compromised dominant preparation in retransition, or a tonal return that coincides with a later module of P, or, potentially TR. The latter point is especially connected to the Type 2 sonata where P consists of several modules. As discussed in Chapter II, the more complicated the expositional structure, the more room for ambiguity around the tonal resolution and accompanying thematic return. In other words, several of the issues that arise around the analysis of Type 2 sonatas 178 specifically about post-developmental space are relevant in Type 3 sonatas that stray from the “textbook” configuration. Hepokoski and Darcy argue, and I concur, that these challenges point to the possibility of formal hybrids between Type 2 and Type 3 sonatas (see Chapter IV) and that this is a rich line of inquiry to be undertaken in future work. Links between formal layout and style—especially in the galant—would also be a worthy contribution to the study of eighteenth-century music. The more we understand the workings of mid- eighteenth-century compositions, the more we can broaden our understanding of eighteenth-century music beyond Haydn, Mozart, and Beethoven. 179 REFERENCES CITED Bach, C.P.E. Essay on the True Art of Playing Keyboard Instruments, trans. William J. Mitchell. New York: W.W. Norton & Company, 1949. Berger, Karol. “The Second-Movement Punctuation Form in Mozart’s Piano Concertos: the Andantino of K. 449,” in Bericht über den Internationalen Mozart- Kongreß Salzburg 1991, Mozart-Jahrbuch 1991, ed. R. Angermüller, D. Berke, U. Hofmann, and W. Rehm (1992): 168–172. Bonds, Mark Evan. Wordless Rhetoric: Musical Form and the Metaphor of Oration. Cambridge, MA: Harvard University Press, 1991. ———. “The Spatial Representation of Musical Form,” Journal of Musicology 27 no. 3 (2010): 265–303. Burstein, L. Poundie. Journeys through Galant Expositions. New York: Oxford University Press, 2020. ———. “The Half Cadence and Other Such Slippery Events,” Music Theory Spectrum 36 no. 2 (2014): 203–227. Caplin, William E. Classical Form: A Theory of Formal Functions for the Instrumental Music of Haydn, Mozart and Beethoven. New York: Oxford University Press, 1998. ———. “Response to the Comments,” in Musical Form, Forms, Formenlehre: Three Methodological Reflections, ed. by Pieter Bergé. Leuven: Leuven University Press, 2010, 51–61. Churgin, Bathia. “Francesco Galeazzi’s Description (1796) of Sonata Form,” Journal of the American Musicological Society 21 (1968): 195–196. Cone, Edward T. Musical Form and Musical Performance. New York: W.W. Norton & Company, 1968. Corneilson, Paul. J.C. Bach, in The Late Eighteenth-Century Composers, series ed. Simon P. Keefe New York: Routledge, 2016. Green, Douglass M. Form in Tonal Music: An Introduction to Analysis. New York: Holt, Rinehart and Winston, Inc., 1965. Heartz, Daniel. Music in European Capitals: The Galant Style. New York: W.W. Norton & Company, 2003. Hepokoski, James A. A Sonata Theory Handbook. New York: Oxford University Press, 2020. 180 Hepokoski, James A. and Warren Darcy. Elements of Sonata Theory: Norms, Types, and Deformations in the Late Eighteenth-Century Sonata. New York: Oxford University Press, 2006. Koch, Heinrich Christoph. Versuch einer Anleitung zur Composition, translated by Nancy Kovaleff Baker. New Haven: Yale University Press, 1983. ———. Musikalisches Lexikon (Frankfurt, 1802), translated by Leonard G. Ratner, Classic Music: Expression, Form, and Style (New York: Schirmer Books, 1980). Hoyt, Peter A. “The Concept of Développement in the Early Nineteenth Century,” 141– 162 in Music Theory and the Age of Romanticism, ed. Ian Bent, Cambridge: Cambridge University Press, 1996. Jackson, Timothy L. “The Tragic Reversed Recapitulation in the German Classical Tradition,” Journal of Music Theory 40, no. 1 (1996): 61–111. Kirkpatrick, Ralph. Domenico Scarlatti, (Princeton: Princeton University Press, 1953. Long, Rebecca J. “Extrinsic Phrases in Early-Classical Sonata Forms,” Ph.D. dissertation University of Massachusetts-Amherst, 2018. Longyear, R.M. “Binary Variants of Early Classic Sonata Form.” Journal of Music Theory 13 (1969): 162-83. Marx, A.B. Die Lehre von von der musikalischen Komposition, practisch-theoretisch, vol. 1. Translated by August H. Wehrhan (London: R. Cocks, 1852). Morrow, Mary Sue and Bathia Churgin. The Eighteenth-Century Symphony. The Symphonic Repertoire 1. Bloomington: Indiana University Press, 2012. Ratner, Leonard. Classic Music: Expression, Form, and Style. New York: Schirmer, 1980. Reicha, Antoine-Joseph. Traité de haute composition musicale. Paris: Zetter, 1824–1826. Richards, Mark. “Beethoven and the Obscured Medial Caesura: A Study in the Transformation of Style,” Music Theory Spectrum 35 no.2 (2013): 166–193. Rosen, Charles. The Classical Style: Haydn, Mozart, Beethoven, expanded ed. New York: W.W. Norton & Company, 1998. ———. Sonata Forms, revised ed. New York: W.W. Norton & Company, 1988. Smith, Peter H. “The Type 2 Sonata in the Nineteenth Century: Two Case Studies from Mendelssohn and Dvořák,” Journal of Music Theory 63, no. 1 (April 2019): 107. 181 Sutcliffe, W. Dean. The Keyboard Sonatas of Domenico Scarlatti. Cambridge: Cambridge University Press, 2008. Vande Moortele, Steven. The Romantic Overture and Musical Form from Rossini to Wagner. Cambridge: Cambridge University Press, 2017. ———. “In Search of Romantic Form,” Music Analysis 32 no. 3, (2013): 408. Warburton, Ernest, compiler. “Thematic Catalogue and Music Supplement,” vol. 48 in The Collected Works of Johann Christian Bach 1735–1782. New York: Garland Publishing, 1999. Webster, James. “Sonata Form.” The New Grove Dictionary of Music and Musicians, 2nd ed., ed. Stanley Sadie and John Tyrell, 23: 687–701. London: Macmillan, 2001. Wingfield, Paul. “Review: Beyond Norms and Deformations: Toward a Theory of Sonata Form as Reception History.” Music Analysis 27, no. 1 (2008): 137–177. Wolf, Eugene K. The Symphonies of Johann Stamitz: A Study in the Formation of the Classical Style. Utrecht: Bohn, Scheltema, & Holkema, 1981. 182