Você está na página 1de 19

Society for Music Theory

Schubert's Experiments with Sonata Form: Formal-Tonal Design versus Underlying Structure Author(s): David Beach Reviewed work(s): Source: Music Theory Spectrum, Vol. 15, No. 1 (Spring, 1993), pp. 1-18 Published by: University of California Press on behalf of the Society for Music Theory Stable URL: http://www.jstor.org/stable/745906 . Accessed: 14/05/2012 01:12
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

University of California Press and Society for Music Theory are collaborating with JSTOR to digitize, preserve and extend access to Music Theory Spectrum.

http://www.jstor.org

Schubert's Experiments with versus Formal-Tonal Design


David Beach
The genesis of this study was an investigation of Schubert's experiments with sonata form, in particular his fondness for beginning the recapitulation in a key other than the tonic, most frequently the subdominant. The purpose of the investigation was not to categorize such experiments, nor to make a catalog of those movements, but to consider how these deviations from classical procedure might affect interpretation of deep-level structure from a Schenkerian perspective.' To provide a proper context for the conclusions reached with regard to this question, it is first necessary to tackle a larger issue; namely, the relationship between what is often referred to as design-a broad category which includes not only surface compositional elements like motive and rhythm, but also features of large-scale organization such

Sonata

Form:

Underlying Structure

An earlier version of this paper was read at the Second International Schenker Symposium held March 27-29, 1992, at the Mannes College of Music in New York City. 'A listing of movements in which the recapitulationbegins in a key other than the tonic can be found in Daniel Coren, "Ambiguity in Schubert'sRecapitulations."Musical Quarterly60 (1974): 568-82. Coren takes a very limited view of sonata form, and thus he excludes from his list several movements that are really sonata forms without developments, that is, a statement of ideas (exposition) followed immediately by a restatement (recapitulation). Some additional information is available in Malcolm Boyd's brief article, "Schubert'sShort Cuts," The Music Reviev 29 (1968): 12-21.

as form and key succession-and underlying structure. This is an enormous topic, far too complex to treat comprehensively in this limited context, yet crucial to an understanding of Schenker's theories and to the interpretation of Schubert's experiments with formal-tonal design offered here. Thus this article can attempt only to introduce this important topic-to raise pertinent issues, to offer possible "solutions," and to suggest directions for further research. The distinction between design and structure is not a new one; it is implicit in Schenker's writings, and the topic has received a fair amount of attention in the literature based on his work. What is lacking in these sources, however, is a consensus about what is meant by these terms and thus about how they are related. A brief summary of this literature is provided in the following paragraphs in preparation for the topic to be developed here. The second part of the essay provides a single example of relationships between design and structure in a brief excerpt, the opening eight bars of Beethoven's Piano Sonata in A-flat Major, Op. 26, and explores some larger issues pertinent to the current investigation. The final section closes the circle. It considers the interaction of formal-tonal design and structure (and the possible effect of the former on the latter) in the initial movements of Schubert's "Trout" Quintet, D. 667, and the C-Major String Quintet, D. 956. Also discussed briefly is the first movement of the G-Major String Quartet, D. 887.

MusicTheory Spectrum

The first to address the interaction of design and structure was Felix Salzer, who distinguished three separate though interrelated factors in musical composition: structure, form, and design.2 The structure of a composition (or passage) may be understood as its underlying voice leading, or, in other words, its contrapuntal-harmonic framework. Form is defined as "a principle of architectonic organization of the structure" (Structural Hearing, 223) and design as "the organization of the composition's motivic, thematic and rhythmic material through which the functions of form and structure are made clear" (224). Here Salzer is following Schenker by defining form in terms of the structure; what is normally understood under the rubric "form" in traditional theory falls, for the most part, under "design." Where Salzer departs from Schenker is in his various divisions of form. First he distinguishes between inner form, the various forms of the detail, and outer form, the organization of the structural progression. The latter category is further divided into structure form, where division, if it occurs, is applied to the structure (for example, a division of the structure into two parts by interruption), and prolongation form, where it is the prolongation of the structure that is divided. Though logical in some ways, this distinction can lead to some confusion. For example, sonata form in the minor mode is described by Salzer as a three-part structure form, whereas sonata form in the major mode is described as a three-part prolongation form. The distinction is made on the basis of where the interruption occurs: in major mode at the end of the first part, in which case 2 is subsequently prolonged, or in minor mode at the end of the second part, just before the restatement. In either case, the three-part form is understood as an elaboration of a two-part division of the structure.
2Felix Salzer, StructuralHearing (New York: Dover, 1962).

A more recent attempt to clarify and elaborate Schenker's brief discussion of form in Free Composition is provided by Allen Cadwallader in his article, "Form and Tonal Process: The Design of Different Structural Levels."3 A careful reading of this essay reveals that Cadwallader's ideas are for the most part a reformulation of what Salzer had to say on the subject. For all practical purposes, the terms "structure," "form," and "design" retain Salzer's meanings, but what is somewhat different is the emphasis placed by Cadwallader on the processes by which form and design are created and his view that "each structural level carries with it its own form and design" (5). By "process" Cadwallader means Schenker's Auskomponierung (unfortunately translated here as "unfolding"), and form is defined as the division/articulation of tonal space by process. Near the end of the essay we are presented with the following tentative generalization about the distinction between form and design. Form refers simply to the basic patterns resulting from the articulative effects of tonal processes on different levels of tonal space. Design-also generatedthroughprocess-is an attendantassociation and is more specific in its application;it signifies the characteristics of the processes, the physiognomy of the patterns.4 Though elegantly stated, Cadwallader is not saying much new here. Also, it is disappointing that more attention is not paid to the relationship of this approach to the more traditional view of form. (It must be kept in mind, however, that Cadwallader's essay is intended only as a preliminary step toward the development of a theory of form from a Schenkerian perspective-a difficult task in itself.)

3Allen Cadwallader, ed., Trends in Schenkerian Research (New York:

Schirmer, 1990), 1-21. Ibid., 15-16.

Schubert'sExperimentswith Sonata Form 3

One individual who has made an effort to reconcile the Schenkerian and the more traditional notions of form is William Rothstein in his recently published book, Phrase
Rhythm in Tonal Music.5 Rothstein recognizes the potential

confusion inherent in Salzer'sterminology and, in particular, Salzer'sdistinctionbetween form and design. Yet he borrows
Salzer's terms outer form and inner form, giving them new

meanings. The former is used to designate "the thematic aspect of the piece, as well as its layout into phrases and periods" (104), in short what Salzer had meant in part by "design." And innerform is redefined as "the tonal dynamic of a work-its large-scaleharmonicand linear layout," which he equates with Salzer's use of the word "form" (although this usage seems closer to what Salzer meant by "structure"). In any case, the distinction Rothstein wants to make between surface phenomena and underlying organization is certainly a valid one; it is, in fact, fundamentalto Schenkerianthought. The common thread through these sources is a general agreement about what is meant by "structure," namely the underlying voice leading and harmonic organization of a given work or passage, which is how the term will be used here. To be more precise, one might use the expression "voice-leading structure"or "tonal structure,"but since this meaning is understood, the term "structure" is sufficient. Unlike other treatments of this topic, however, the present discussionincludes all other aspects of organization,including form, under the general heading of "design," but because this category embraces many types of organization, consistency is demanded in the use of modifiers, unless one and only one type of design is under discussion. One might speak, for example, of the "formal design" of a composition, referringto its division into sections, which may be subdivided perhaps into "themes" and connecting passages, then periods and
5WilliamRothstein, PhraseRhythmin TonalMusic (New York: Schirmer, 1989).

phrases, and so forth-in short the traditional hierarchical notion of form (as opposed to Schenker's notion of form as a product of voice leading). Correspondingto this aspect of organization is "tonal design," namely the layout of keys in the course of a composition, which is an entirely different matter than tonal structure. (We will returnto this important distinction later.) Thus, at the macro level, distinctions should be drawn between underlying structure and formal design on the one hand, and between structure and tonal design on the other. Also significant, of course, is the distinction between formal and tonal design, which often, but not always, correspond. There are many other aspects of design one might consider-for instance motivic design, registral design, rhythmic design, and surface design. Each of these areas is worthy of extensive study in itself, and a thorough investigation would include not only interactions between structure and design but also among different aspects of design.6 Clearly such an investigation is beyond the scope of this study. Instead the focus here will be primarilyon the relationship of large-scale formal and tonal design to underlying structure. One reason for the approach taken here is to facilitate comparisonsbetween different conceptions of musical form.

6An importantarticle that deals with the interactionof design and structure at more immediatelevels is John Rothgeb's "Design as a Key to Structure in Tonal Music," Journal of Music Theory 15 (1971): 230-53. In this study Rothgeb demonstrates that changes in surface design or pattern often correspond to crucial structuralpoints. Viewed from another perspective, such changes can aid us in making decisions about or help us verify our interpretation of underlying structure. Another articledeservingof mention in this context is Ernst Oster's "Register and Large-ScaleConnection," Journalof Music Theory5 (1961): 54-71. In this study, Oster shows the important role register (an aspect of design) can play in articulatingstructuralor motivic connections. Sometimes these connections transcendmajor formal boundaries, thus creating "conflicts"between registral and formal design.

MusicTheory Spectrum

Schenker, of course, was not much interested in accommodating the traditional notion of form, and understandably so. He was interested in presenting a new theory-or at least the seeds for a new theory-of form based on his new conception of musical structure. There might be some today who will argue that nothing has really changed in this regard during the last fifty-some years, but there are at least two compelling reasons why we ought not adopt such an attitude today. First, it is very important that we develop a vocabulary that will facilitate communication between Schenkerians and othersare making musicologists, performers and theorists-who worthwhile contributions in this area. Second, in this author's opinion there is some music (some late instrumental works of Schubert come to mind) that is best understood through a combination of approaches. This may be true for all music, but a combination of approaches seems particularly apt for much music of the nineteenth century. In the preceding discussion reference was made to the difference between the Schenkerian and the more traditional notions of form. To illustrate, let us take sonata form as an example. The traditional view (setting aside the early descriptions from the late eighteenth and early nineteenth centuries) is that the macro-form is ternary, consisting of the familiar divisions into exposition, development, and recapitulation, with various subdivisions into primary and secondary themes, and so forth. Schenker also understood the form as ternary, but for very different reasons. In his view, the first section of a sonata movement in the major mode is defined by a motion to 2, the second is a working-out of the dominant, often involving the introduction of the seventh (for instance V5-7), and the third, the restatement, completes the motion to closure. As noted before, the origin of this "form" lies in a two-part division of the structure, the first part of which has been extended by an elaborate working-out of the dominant. The extension is indeed a formal unit, but hardly equivalent to the other two, and thus it seems more logical-and, in fact,

more consistent with Schenker's view of structure-to consider sonata form as fundamentally binary in nature and perhaps ternary only at some lower level. One might say it is binary in structure, but ternary in design. This brings us back to the original topic: establishing a common vocabulary. The term "formal design" has already been suggested to describe the traditional hierarchical nature of form. And since the structural view is tied to the notion of interruption, it seems logical, on the other hand, to speak of a two-part structure, or of parts one and two of the structure.
w. w.

Before delving into some larger-scale issues, the relationformal and motivic design-to ship of design-specifically structure shall be examined in one brief excerpt, the opening eight measures of Beethoven's Piano Sonata in A-flat Major, Op. 26.7 These eight bars, which are reproduced at level c of Example 1, constitute the antecedent phrase of a sixteenbar parallel period, the formal design of which is diagramed in Figure 1. This diagram shows that each eight-bar phrase is divided into two four-bar subphrases (for lack of a better term), and the letters x and y indicate that these subphrases are contrasting, as opposed to similar or parallel, in character. The punctuation symbols separating the roman numerals indicate a hierarchy of goals, the most important distinction being that the dominant falling at the end of the antecedent phrase has a higher status than those articulating the internal divisions of the phrases. This distinction may seem obvious, but is important to make, since interpretation of the structure
7A detailed discussion of this theme is provided in David Beach, "The

Analytic Process: A Practical Demonstration," Jolrnal of Music Theory Pedagogy 3, no. 1 (Spring 1989): 25-46. See also Figure 85 in Heinrich Schenker, Free Composition, trans. and ed. Ernst Oster (New York: Longman, 1979).

Schubert's Experimentswith Sonata Form 5

Example 1. Beethoven, Piano Sonata in A-flat Major, Op. 26, i, mm. 1-8
A

3 3rd

3rd

//

l-----

4bbkb
r^

;
I

^-f^
p.t. cons.

'

I ____/

IV6 \

i^bJbbe

'"-~

J
P

L-

J V.^ ^
A A4

: e:
AA A A

J
,p,

J
l

r':^Sbb
I

4 ,

_
6

"^T

j-'
4 V, v , IV6 2

A -7

depends very much on how we view the roles of these dominants. This interpretation says that the dominant in m. 4 articulates a phrase division and that not it, but the dominant at the end of the phrase, is the real goal of tonal motion. The

alternative, of course, is to interpret the contents of mm. 5-8 as prolonging the dominant from m. 4. Taken in the abstract, this is certainly possible, but in this context it makes little musical sense. It is difficult to hear, let alone play, the phrase

MusicTheory Spectrum next four measures grows. Likewise, of course, one will miss this connection entirely if no heed is taken of the voice leading. The point is this: only a partial understandingof this phrasewill result from consideringeither surfacearticulations or underlyingvoice leading, either formaldesign or structure. There are many importantand interestingissues one might consider under the general heading of "design-structure," but, as noted above, we cannot attempt to be comprehensive in this limited context. One topic we shall pursue relates to an issue this authorraised a few years ago in an article entitled "A RecurringPattern in Mozart'sMusic."9In that study several examples were discussed where the goal of the tonal motion in the development section is the major triad on the mediant (III"), after which the tonic harmony, in conjunction with the opening material, is restated either directly or througha passing dominant, which is sometimes extended for several measures. In the former case, the progression connecting the exposition to the recapitulation is V-III-I, as shown in Example 2. In the particularmovements discussed, the descendingfifth of the tonic triad is a primarymotive, and thus the point was made that the events of the development section might be viewed as an elaborate composing-outof this motivic idea. In short, like the descending third db-c-bl in the Beethoven example, the surfacemotivic idea here too has given rise to an extended passage; in the Beethoven it was the next four bars; here it is the entire passage connecting the exposition and the recapitulation. But the principle remains the same. In the latter situation-where the connection to the tonic is accomplished through a passing dominant-the inwhen that dominant terpretationis not so simple, particularly is extended for several measures. One is faced with two possibilities, shown at a and b in Example 3. Where the connecting dominantis brief, as in the firstmovement of Mozart's
9DavidBeach, "A RecurringPatternin Mozart'sMusic,"Journalof Music Theory 27 (1983): 1-29.

Figure 1. Beethoven, Piano Sonata in A-flat Major, Op. 26, i, mm. 1-8, formal design
Antecedent Phrase Consequent Phrase

/
/ I

X
4 \1 V, IV6

Y
4

\
\ V;

X1
4 I \ V, IV6

Y
4 V

\
\ I

in mm. 5-8 as if it were extending a goal already achieved. This interpretationis counter to the musical idea of playing throughthe internalphrase division to the end of the phrase.8 The structuralfunction of the initial dominant (m. 4) in relation to the voice leading is shown at level a of Example dominant prolong1. It is interpreted as a backward-relating ing the initial tonic and supporting a passing tone between the primarytone c2 (3) and the subsequent inner-voice tone ab1. The slurs above the treble show that the initial ascending
third a b-bb -c2, which occurs within the prolonged tonic,

is answered by the descending third db2-c2-bb , the final pitch of which is supported by the local structuraldominant. Thus the underlying structure of this two-part phrase is the division 2 . One additional feature of this passage should be mentioned: the development of the motivic third db2-c2-b
1,

occurrences of which are marked in the graph at level b by square brackets. The initial statement of this idea occurs on the downbeat of m. 4, where it receives strong rhythmic articulation. Here the db is an appoggiaturaand is thus not part of the voice leading, but one must be careful in looking for underlying structure not to overlook this important idea, since it is the seed from which the melodic material of the
8In this author's opinion, many important decisions about underlying structure are tied directly to musical understanding,not to theoretical justification.

Schubert's Experimentswith Sonata Form 7

Example 2. Recurringprogressionconnecting exposition and in recapitulations Mozart'ssonata forms


r
#6
V IIIt I

of Example 3. Two interpretations the connectingdominant


a or b 5-7

1>
6 ?

^ 3

4
V

44

r
I

Piano Sonata in F Major, K. 280, the reading at a seems to apply. But when the connecting dominant is extended, as in the first movement of the Sonata in F Major, K. 332 (even more so in the first movement of the B-flat Major Sonata, K. 333), then the interpretation at b seems more appropriate. That is, despite the compositional emphasis given to the major triad on the mediant, we understand that harmony as functioningwithin a prolonged dominant, perhaps as support for a passing tone connecting the fifth to the seventh. What has just been described might be called-to use Carl Schachter's expression-an Either/Or situation, where a choice must be made between alternativereadings after careful consideration of the evidence.10 It is certainly true that informed choices are made at various stages in the analytic process, but there are situations like this-where the compositional articulation suggests one thing, while the underlying progression suggests another-where it seems counter'0CarlSchachter, "Either/Or," SchenkerStudies, ed. Hedi Siegel (Cambridge: CambridgeUniversity Press, 1990), 165-79.

intuitive to eliminate one in favor of the other. Echoing a question posed earlier by Schachter, one might well ask: Is it possible that both interpretationsare valid? In the opinion of this author, the answer is yes, as long as we understand that the two represent different aspects of the composition. One (a) representsthe articulatedtonal design, which, as has been pointed out, might be related to the motivic design; and the other (b) represents the underlying structure. In fact, a convincing argument could be made that the structure of Example 2, where there is no connecting dominant, is also V->I. The example just presented raises at least three important questions. First, just what is it we are trying to represent in an analyticgraph?The answer, if we are talking about Schenkerian analysis, is structure. This does not at all imply that Schenker was disinterested in other aspects of musical organization; this interest in many facets of design is evident in his extensive commentaries. In fact, even in his graphs he often went out of his way to indicate important features of design, such as motivic parallels. But the fact remains that Schenker's graphs, at least his later ones, represent his interpretation of structure. Second, to what extent must we choose between alternative readings? If we are talking just about structure,then this authorconcurswith Carl Schachter: We must make choices based on the evidence. (In fact, later in this article preferred choices will be specified when there seem to be viable alternatives of interpretingthe structure.) But if we are willing to accommodate aspects of tonal design as well as tonal structure,then we are not forced to eliminate one in favor of the other. We are most certainly obligated, however, to make it perfectly clear which we are talking about at any given time. Finally, if we are to accommodate both design and structure, how might we notate one versus the other to make the distinction clear? A reasonable approach would be to use one notation in design and an entirely different notation in a sketch showing structuralconnections,

MusicTheory Spectrum

Example 4. Two interpretations of the connecting dominant represented simultaneously

Example 5. Brahms, Symphony in D Major, Op. 73, i


A

6 r,
j6

J
#?
3

X~ (f)
r

A f _L
a

,II -

and, in addition, to represent each separately. The two can be represented simultaneously, as in Example 4, where the tonal design is represented by a beamed connection and the structure by a slur. In such graphs it is of course absolutely necessary to make clear just what is representing what. The very distinction drawn here between tonal design and tonal structure is at the heart of Carl Schachter's discussion of the path leading from the tonic to the dominant in the first movement of Brahms's Second Symphony." The following is his description of the key scheme in relation to the formal plan-or, in the terminology favored here, the tonal design. Like some of Schubert's expositions (and others by Brahms), this one divides into three main key areas instead of the usual two. The 'second theme' appears in the mediant key, F# minor. A brief transition follows, culminating in an augmented sixth that leads to a strong V of A major, the dominant key. The remainder of the exposition is in A.12 Schachter goes on to say that most analysts would interpret the large-scale tonal motion as I-iii-V, as shown by the graph at level b of Example 5, which represents the articulated tonal design. But Schachter does not believe this interpretation is
"Carl Schachter, "The First Movement of Brahms's Second Symphony: The First Theme and its Consequences," Music Analysis 2, no. 1 (1983):
55-68.

II

(UI#)

111

12Ibid.,62.

correct for reasons he outlines in his article. Instead he provides evidence that the structure-what he calls the tonal an extended chromaticized voice exchange, plan-involves which extends the tonic harmony beyond the boundary of the second theme area (in F-sharp minor) before leading to the dominant. In short, he interprets the structure as I-(IIt)-V. In all likelihood he is right; that is the underlying structure. The implication of Schachter's choice, however, seems to be that one possibility is being eliminated in favor of the other. In one sense, as a decision about the structure, that is true. But it may reasonably be argued-and perhaps Carl Schachter would agree-that an understanding of this movement and many others like it depends on much more than the proper identification of structure. His decision is based on considerable understanding of many levels and types of

Schubert'sExperimentswith Sonata Form 9 organization, including the articulated formal-tonal plan I-iii-V. It is important to recognize both means of organization and their interaction, just as it is crucialnot to confuse the one for the other.

Example6. Schubert,Quintetin A Major,D. 667 ("Trout"),ii

(
Schubert'sinstrumentalmusic provides fertile territoryfor investigatingthe interaction of formal-tonaldesign and tonal structure. The discussion here will focus on but one feature of Schubert'sstyle-his frequent practice of preservingin the recapitulationthe tonal scheme of the exposition, thus making the tonic the goal rather than the point of departure in the restatement. The most common pattern is I--V answered by IV--I, the well-known subdominant recapitulation. But sometimes-particularly when the dominant is no longer the goal of tonal motion in the first part-the result is a rather odd tonal scheme, as in the second movement of the "Trout" Quintet. Example 6, a diagram of the key succession of this movement in relation to three clearly articulated thematic ideas (indicated by circled arabic numerals), shows that the progression F major-Fit minor-D major is answered by Ab major-A minor-F major. One might argue for various reasons, formal and tonal, that this movement is not a real sonata form. In fact it does not matterwhat label one applies, as long as there is agreement that it is organized into two large parts-an antecedent (statement of ideas) and a consequent (a restatement transposed to end on the tonic). It should be clear from looking at this tonal scheme that there can be no large-scale interruption, since there is nothing to support 2, except perhaps G major as substitute for the dominant. But G major has only local significance here, as IV in D major and as a lead-in to the following key of A , major. In this case, the tonal scheme has affected the underlyingstructure.There is no interruption;instead the primarytone (3) is prolonged until the return of the tonic (idea 3) in the restatement.

0(

C0 00

r
keys: \F

:r 7 ttr
(I ft D! IV) \Ab a~ F/

One could point to several examples of unusual tonal schemes in Schubert's instrumental music, but it is not the intent of this study to provide a complete listing of these works. Instead the following discussion will focus on the initial movements of the Quintet in A Major, D. 667 ("Trout"), and the C-MajorStringQuintet, D. 956, from the perspective of the interaction of tonal design and structure, with special attention to the potential effect of the former on the latter. Certain features of the first movement of the G-major Quartet, D. 887, will also be discussed briefly. The first movement of the Trout Quintet provides a clear example of the subdominant recapitulation. Graphic representations of the voice-leading connections in the exposition of this movement (up to the closing material) and the corresponding passages in the recapitulation are provided in Example 7, which is organized to show that the latter is a transposition of the former down a fifth. Following the introduction (mm. 1-24), which contains the primarymotivic material of the movement, three clearly differentiated ideas are presented, each of which is repeated in slightly varied form. The first of these is an outgrowth of the introduction,

10

MusicTheory Spectrum

Example 7. Schubert, Trout Quintet, i


Intro. Th. 1
A

Th. 2
A

Th. 3 (*)

2 5th

"~~__ --

__1

___" 1

a ..~~~~~~~~~,i

v
N 5th

IK

278/285 Recapitulation 10) l


4
''):#WH

(2 c B

( 69 A)

etc.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i-i?
7 h6 4 6 5 18 6 4 b7 5 3 6 5 V7 6 5

I~~~~~~~~~~~~~~~~~~~
I

6 4

5 3

IV *See Free Compositiont, Fig. 109b

and the graph represents the voice leading of the varied repetition. As shown between the staves, the underlying melodic motion is based on the represented melodic content of the The primary feature of this introduction: Ct-C -C#-B-A. idea, which acts in effect as a basic motive, is the modal change Ct-Cb, followed by a brief recapturing of the Ct before the descent to local closure. For the second theme,

which is characterized by a dialogue between cello and violin, features of the original statement and the varied repetition have purposely been combined to show that an articulated inner line is a transposition of this motive up a fifth. Thus, when this theme is stated in the tonic key in the recapitulation, the motive occurs at the original pitch level. Finally, the third theme, first stated by the solo piano, outlines a

Schubert'sExperimentswith Sonata Form 11

descending fifth, a gesture that is repeated several times before the close of the exposition.13 The primarytonal motion of the exposition is I to V supporting 3 to 2, which is prolonged by the fifth-descent of Theme 3. Interpretationof the recapitulationis not so simple, due to the subdominantreturn. Example 8 provides two possibilities, labelled a and b. In the first of these, the subdominant is interpreted as existing within a prolonged dominant extending from the beginning of Theme 2 (m. 64) throughthe dominant that introduces the same idea in the recapitulation
(m. 248).14 A simplification of the voice leading is provided

below in brackets, which shows that the subdominant gives temporary consonant support to what in the larger context is the seventh of the covering 8-7 motion that reintroduces the primary tone Ct. This interpretation preserves the interruption of the fundamental structure, which is basic to Schenker's conception of sonata form. The fundamentalline then descends to closure through the motivic inner voice of the second theme, the importance of which was stressed above, and the covering fifth is resolved by the descent of the third theme. Though this interpretation of the structure makes sense from several points of view, it depends on the following assumptions: 1) that the dominant in m. 248 is sufficientto recall the structuraldominant, and 2) that we can hear closure in m. 269. The second option denies both, and, in so doing, suggests that the tonal design has altered the structure. It says, in essence, that there is no division of the
'3ThoughThemes 2 and 3 are distinctideas by virtue of their articulations, careful examination reveals that the third is derived from the second. Thus, in a sense, there are just two underlyingideas, correspondingto the two main tonal areas. '4This interpretation is similar to Edward Laufer's reading of the first movement of Schubert'sFifth Symphony. See p. 36 in Laufer's handout for "Voice-LeadingProceduresin Development Sections," a lecture given at the conference Music Theory Canada 1990, held at the University of Western Ontario.

structure,but ratherthat the tonal and thematic design of the recapitulation delays closure, which comes only at the conclusion of the third idea. If one were to accept the validity of this interpretation, would this mean that this movement is not a real sonata form? If form is considered an attribute of design, not of structure, then there is no problem in identifying the work as a sonata form. The formal design remainsintact, but the structureis a single undividedmotion. From the standpoint of either option a or b, there is a clear distinction between the formal design and the underlying structure. The existence of two possible interpretationsof the structure raises an importantissue, namely that of making analytic choices. Earlier it was suggested that two readings of a passage or movement might be valid, one of the tonal design and the other of the structure, as long as the distinction was clearly drawn. (One might logically extend that idea to several readings, depending upon the number of variablesbeing considered in isolation or in combination, though this author is personallyreluctantto do so, preferringto think of analysis as a synthesis of information rather than a division into a set of variables.) But when it comes to a consideration of structure, that is, the voice leading at various levels, Carl Schachter's Either/Or paradigm is unquestionably sound. The analyst must first understand what the possibilities are and then make informed choices based on careful consideration of the evidence. Given the meaning of the term structure as it is used here (Schenker's notion of structure), the possible interpretationseventuallylead to one (althougheven Schenker was known to change his mind!), no matter how difficult the choice may be. Returning to the two possible interpretationsof the initial movement of the Trout Quintet given in Example 8, three reasons may be stated why the first of these makes more sense: 1) it provides a structural interpretation for the subdominantreturn (consonant support for the seventh of the controlling 8-7 motion over the

12

MusicTheory Spectrum

Example 8. Schubert, Trout Quintet, i: Two interpretations


A

Recapitulation

2 ]t 7r~~~~~~~~~~~

5th

3 2
(1) (2) (3) Dev. (1) (2) (3)

"? #ttit

_7

\ ,

,
77

/-.
V I

r
1 V V I

8-

(8)

-7

(1)(2)

(3

________

Dev___

(a

-3

.rI

?
V

4-

.. IV IV

Schubert'sExperimentswith Sonata Form 13

dominant); 2) it shows very clearly the distinction between the formal design and the underlyingdivision of the structure; and 3) it shows the subdominant return as a variant of an underlying structural norm, which is how this author hears it. While these points provide support for choosing option a, it must be pointed out that they do not invalidatethe alternate reading given at b. Some of the same issues encountered with the first movement of the Trout Quintet, plus several new ones, arise when we examine the relationshipof design to underlyingstructure in the firstmovement of the C-MajorQuintet. A graphshowing the main features of the exposition and the recapitulation of this movement is provided in Example 9, which, like Example 7, is organized to show that the latter part, beginning with the initial thematic idea (1), is a transposition, down a fifth, of the correspondingpassages from the exposition. Before discussing matters pertaining to interpretation of the structure, certain features of this movement represented in the graph should be noted. First, Example 9 contains various references to the chromatic inflection of scale-degree three (e-eb/d#-et), an idea that has more than local significance in this movement. In the opening measures the ek is introduced as part of an embellishing diminished-seventhchord, and later it appears-or rather is implied-as a dt, the third of a B-major chord, which functions as IIIt of the dominant. (Chromatic mediant chords function at several levels in this movement as extensions of the local controlling harmonies.) The harmonizationof this motive in the first "theme," which is an outgrowth of the introduction, is altered to direct the motion more strongly to the supertonic. And in the third "theme," d# and later eb appear as decorations of dO(2). In the former case, d# appears once again as the third of a B-major chord, III# of the dominant, and eb is stated twice in the cadential progression II-V-I confirming the modulation to V. All these occurrences, which are representative but by no means exhaustive, are notated in lower-case letters.

An enlarged statement of this idea, marked in the graph by upper-case letters, spans what are marked as the first and second themes. The latter of these in particularpresents several problems of interpretationto the analyst because of the references in it to earlier/later harmonies. In the local context, the controlling tonality is Eb (bIII), and while the G-major chord may be taken as a reference to the dominant, its function here is as the upper third (III) of Eb. (At the correspondingplace in the recapitulation, the potential ambiguity of the C-major chord is even greater: locally it is III1 in Ab, but it also refers back to C as the dominant of IV as well as forward to the tonic itself!) The instability of this second key area has been noted by Charles Rosen, who recognizes that the key of El in the exposition is an extension of the tonic. 15 Not only does the return to the tonic harmony in m. 71 (or later again in m. 92) signal the end of the tonic prolongation and the simultaneous departure for the key of the dominant, but it also corresponds to the return of E~. Thus, while the tonal-formal design of this movement articulates the progression I-bIII-V, the structuralpath to the dominant is not through the mediant. Rather, from a structural perspective, Eb major functions as the upper third of the tonic and, at the same time, as support for the Eb within the expanded statement of the motive EB-Eb-EB. An important difference between this movement and the first movement of the Trout Quintet with respect to largerscale formal design is the restatement of the introductory passage at the beginning of the recapitulation. How we hear this passage in relation to the whole is crucial to how we interpret the fundamental structure. That is, if we hear it as a real return to the tonic, supportinga return to the primary tone (3), then the underlying structure would be something like that shown in Example lOa. This interpretation corresponds to Schenker's concept of sonata form as an
S1Charles Rosen, Sonata Forms (New York: Norton, 1980), 244-45.

Sb#_bS

"

St,li

14

MusicTheory Spectrum

Example9. Schubert,Quintet in C Major, D. 956, i

(D
Intr ro.
A

Q
A

3 . I_ i e/1;

2 N 5th -i7 \
'4);

-'<

/--

>

____1 I
-

:)eb 0~

b-o

-,
$L

K 1-e
-

_b*y=.-' **9

-f

,"

q-R*v

?' #d E/

b--

$
(d eb d) etc.

(eh eb e4)

Etl

/
/0

Eb .-

El

(d d# dd)

6 ii I

56

4
I

.6

(V

%
\^__^
A

0/

4s ff

6 ^

6
5th

6 ^

87
4

Intro,

6
(ei eb e0) (e d4 e) 295

--,--.

11.11, ^

i1 'R t7
etc.

?@~~
r
-r

(~~285 5 :6 5 #6 (#) 4 -6.--4 2

311

322/343

362(400

2Tql c

r F
V

7 _. _ --" b[6
6 4

7
5 3

(I)

IV

expansion of a two-part division of the structure. However, if we hear the motion continuing through this passage to the subdominant, as shown in Example lOb, which is a simplification of Example 9, then there is no interruption. Rather, the structureis undivided, and the motion to the subdominant is shown as occurringwithin a prolonged dominant extending from the third theme of the exposition up to the introduction

of the same idea in the recapitulation. It is difficult to make a choice between these two interpretations, since it seems fairly easy to hear the music either way. The first has the advantage of showing the motion to the subdominantin relation to the structuralnorm, but the second seems to reflect more accurately the passing function of the tonic at the beginning of the recapitulation.

Schubert's Experiments with Sonata Form 15

Example 10. Schubert,C-MajorQuintet, i: Two interpretations


A

IV

An investigationinto the issue of tonic returnsthat are not real returnsin a structuralsense, and thus are somehow passing in function, is a subject for a separate study. The topic came up briefly in a conversation several years ago between the present author and the late Ernst Oster in which we discussedthe second movement of the Unfinished Symphony. The form of that movement is a large two-part design, an exposition containingtwo distinct themes, the firstin E major (I) and the second in Cf minor (vi), followed directly by a

restatement, where the keys for the two themes are E major (I) and A major (IV). There is no dominant leading to the restatement of the initial theme in the tonic, and thus no division of the structure. As shown in Example 11, which is based on Oster's comments, the function of this tonic is to lead to IV, which completes a descending arpeggiation throughvi. It was in reference to this type of situation-where the tonic return functions as the dominant of IV-that the first movement of the C-Major Quintet was brought into the

16

MusicTheory Spectrum recapitulation, as if to begin the process of resolving this conflict. (The subsequent motion to IV might also be understood as part of this process.) Following the introductory phrase are three varied statements of an idea labeled Theme 1 for lack of a more appropriate term. Though the bass line contains all the pitches of the chromatic descent from the tonic to the dominant, the underlyingpattern is the descending tetrachord of the minor mode, supportinga succession of parallel triads. The detail of the voice leading somewhat mitigates this fault, though the repetition of surface patterns reinforces the listener's awarenessof these parallels. The third statement of this idea is extended, leading to the close in m. 54, which is simultaneously the point of departurefor the second theme. (Each step of the ascending fifth sequence connecting the G major and F# major harmonies shown in Example 12 involves direct modal mixture!) There are several unususal features of this second theme. First, it does not begin upon arrival at the new key (the dominant), but rather leads to it. Second, there is the somewhat odd tonal syntax of the theme, which might be summarized by the progression III#-V-I. The bass line articulates the third FO-E-D, which, when considered in relation to its point of origin, is part of the descendingtetrachordfrom I to V in the major mode. Thus, the earlier descent G-FbEk-D has been corrected, so to speak, to G-Ft-E~-D, an expression of modal mixture at the middlegroundlevel. This plausibly explains the emphasis given to the Ft-major triad as the point of departure for this second theme. Finally, the theme is stated in varied form four times in succession, twice connecting Ft to D, then D to B,, and (following a transition to F,) finally leading to D, completing the prolongation of the dominant by a chain of major thirds. With minor alterations, one being the reversal of the modal shift already noted, the restatement of the opening material in the recapitulation is as before, firmly reestab-

Example 11. Schubert,Symphonyin B Minor, D. 759 ("Unfinished"),ii

rC2I ^_ e 8-7

i2 l

T
0. ) :^
1

j
-

^ "-

17

Ir

conversation. Oster specifically mentioned the idea that the arpeggiation C-Et,-G (exposition) is answered in the recapitulation by the arpeggiation F-Ab-C. As already noted, this author attributesthat scheme to design ratherthan structure, and that, presumably, is what Oster meant as well. Whetherconsidered an aspect of design or structure,this view is consistent with the notion of the tonic return as passing-as a lead-in to IV. There are several movements from the late Schubertworks which incorporatethe idea of the subdominantreturninto the traditional tonal scheme of sonata form, in which, in other words, the subdominantprolongs the structuralreturn of the tonic. A clear example of such a plan is the first movement of the last string quartet, a sketch of which is provided in Example 12. (This sketch shows only the most general connections, and thus, by necessity, excludes much interesting detail.) One of Schubert's favorite devices, modal mixture, is nowhere more evident than in this movement; it pervades the detail and controls larger gestures. The change from B~ to Bb is stated in the dramaticopening gesture of the movement, and Schubert reverses the order at the opening of the

Schubert'sExperimentswith Sonata Form 17 Example 12. Schubert,Quartet in G Major, D. 887, i


Th. 1
A

trans.

Th. 2

n(

__

t+

--

'

f
(54) 8

lI-81 5*

'T'
(77/90) (122) (141) (154)

(64)

5'

5__

dIp-

~~~~~~IIIIt

~
Th. 2

7
V3

V v

Th. 1
A

trans.
A

4 A

r3 |

(278...) 55 5 5!

(333)

(342)

(369) 8 5

(388) 8, 5

(401)

b <r x

mI i

hr i^'6

'

*
5

r I I-1I_ IIIt
V4 IV

r I I 111
V4

I I

r
V V

a>

lishing the tonic and confirming D (5) as the primary tone. Rather than leading to a close in the tonic, however, as this section did in the exposition, the extended third statement of the first "theme" leads to a cadence on IV. If the following material were to proceed as a transposition of the corresponding passage in the exposition, as might be expected from Schubert, it would lead to a B-major chord, and from there to closure at the end of the initial statement of the

second theme. Instead Schubert proceeds to an E-major harmony (IIIf in the key of the subdominant), so that the initial statement of the second idea leads to a strong cadence again on IV, thus prolongingthat harmony and avoiding premature closure. Then he proceeds to the B-major harmony, and from there the second and third statements (there are only three in the recapitulation!)lead to the tonic and closure. As shown in Example 12, the extended subdominant is

18

MusicTheory Spectrum

interpreted as providing support for scale-degree 4 of the fundamental line. There appears to be no interruption of the fundamental line; rather, this movement is an example of a sonata form in which 5 is prolonged until the structural descent in the recapitulation.16

The fact that Schenker did not publish analyses of any of the Schubert works cited in this study is intriguing. Though we shall never know the reason, it may be speculated that they (and other works like them) presented a dilemma for Schenker. Assuming he might have perceived that at least some of them do not conform to his conception of sonata form as an outgrowth of a two-part division of the fundamental structure, he would have been faced with two alternatives: 1) to reconsider his ideas on form, which were never fully developed, or 2) to judge the Schubert works as inferior (say, to Beethoven's) in some respects. Schenker most certainly was not shy about using his theories as a yardstick against which to measure other theories or musical compositions, but he may well have avoided criticizing works he probably admired simply by not writing about them. This would explain his silence. The problem would appear to lie

with Schenker's conception of form, which is too restricted to accommodate the types of situations raised in this study. This is what has led this author to take a more traditional approach to the subject-to view form as an aspect of design rather than an outgrowth of structure. There are at least three distinct advantages to this approach. First, one can accommodate both a structural interpretation of the voice leading and the traditional notion of form, which Schenker had rejected, without compromising the integrity of either. And the two can be compared. Second, in both cases norms and deviations from those norms can be discussed. And finally, one can consider the possible effects of deviations in tonal design-for example, the subdominant recapitulation-on the underlying structure. ABSTRACT The basic idea of this article is to investigate how Schubert's experimentswith sonata form-particularly his fondness for beginning the recapitulationin the subdominant-might affect interpretation of deep-level structurefrom a Schenkerianperspective. To explain Schenker's views and to place the discussion in a larger theoretical context, the author first considers the importantdistinctionbetween aspects of musical design and tonal structurein general and then as it relates to our understandingof sonata form. Schenkerconsidered sonata form an outgrowthof a two-partdivision of the structure,but the author prefers to view form as an aspect of design separate from the structure, and he contends that deviations from the classical norm, as exemplified by some of Schubert'smusic, are best understood with this distinction in mind.

16SeeErnst Oster's extensive footnote on pp. 139-41 of Schenker's Free Composition (New York: Longman, 1979).

Você também pode gostar