Jump to ContentJump to Main Navigation
Bach's Cycle, Mozart's ArrowAn Essay on the Origins of Musical Modernity$

Karol Berger

Print publication date: 2007

Print ISBN-13: 9780520250918

Published to California Scholarship Online: May 2012

DOI: 10.1525/california/9780520250918.001.0001

Show Summary Details
Page of

PRINTED FROM CALIFORNIA SCHOLARSHIP ONLINE (www.california.universitypressscholarship.com). (c) Copyright University of California Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in CALSO for personal use (for details see http://california.universitypressscholarship.com/page/535/privacy-policy-and-legal-notice). Subscriber: null; date: 21 June 2018

Mozart at Play

Mozart at Play

(p.179) 4 Mozart at Play
Bach's Cycle, Mozart's Arrow

Karol Berger

University of California Press

Abstract and Keywords

This chapter starts with Mozart's instrumental music. At some point between the early and late eighteenth century, between Bach and Mozart, musical form became primarily temporal, and the center of attention for musicians shifted toward the temporal disposition of events. The discussion demonstrates this point by considering the temporality of the instrumental genre that was of central importance to Mozart, the concerto allegro. Among Mozart's instrumental genres, the concerto stands out for its profusion of melodic invention. In these concerto allegros, the main focus of interest is the order in which melodic ideas are presented within each of the three tellings of the story—opening ritornello, exposition, and recapitulation—making each of the tellings progressively more complete and logical.

Keywords:   instrumental music, Bach, Mozart, concerto allegro, melodic invention

The concept of form involves the interrelated concepts of the whole and its parts. Only an object that is a whole and articulated into distinct parts can be said to possess form. Form is an intelligible relationship between parts and a whole, where all the parts, rather than being merely a heap of unrelated elements, contribute to the establishment of the object as a whole. It is not necessary but certainly most natural for the parts to be organized hierarchically: just as an object may enter into intelligible relationships with other objects by becoming a part of a larger whole, so may a part be articulated into its own parts and become a lesser whole in its own right. To understand the form of an object is to understand how it is divided into parts and how the parts are related to one another and to the whole—that is, what function each part has in the make-up of the whole.

One of my central claims in this book is that at some point between the early and late eighteenth century, between Bach and Mozart, musical form became primarily temporal and the attention of musicians—composers, performers, and listeners alike—shifted toward the temporal disposition of events.1 By Mozart's time, the form of a musical work is temporal; that is, it consists of a number of phases or parts that succeed one another in a determined order. To understand such a form requires recognizing how the object is divided into successive phases and how the phases are related to one another and to the whole—that is, recognizing the function each phase has in the whole, in the transformation of a mere succession of unrelated elements into a configuration of intelligibly related phases.2

For successive phases to relate intelligibly, and a succession to become a configuration, earlier phases must not only precede but also in some way cause the appearance of later ones, and these for their part must not only succeed but also follow from the earlier ones—one-after-another must become (p.180) one-because-of-the-other. For this to happen, the phases must function as beginning, middle, or end. The beginning phase is the only one that does not require any earlier phase, but it does require and in some way must cause the appearance of another, later phase. The ending phase, conversely, does not require a subsequent one but it must require and in some way follow from an earlier one. And the middle phase requires and follows from an earlier phase as well as requiring and causing a later one. Relationships of causing or following, if they are present, may but do not have to obtain between immediately adjacent phases only. More complex and sophisticated temporal forms exhibit a large number of nonimmediate, long-range causative relationships between phases.

The concept of form is inextricably linked with that of matter or material. Form in an object is embodied or materialized; matter is formed. And matter is needed if a form is to be realized. The two fundamental functions of matter are those of articulating the whole into parts and integrating the parts into a whole. If it is true, as I claim here, that the guiding ambition of the classical instrumental composers was to create works in which individual phases do not merely succeed one another but are configured into wholes, the question arises: what were the material means with which they realized this ambition?

Current thinking about classical instrumental music privileges two factors—key and theme—as particularly important for the creation of forms.3 In this view, late eighteenth-century musical logic primarily concerns a tonal plan based on opposition between stable and unstable tonal areas; of secondary importance is a superimposed thematic plan based on exposition, development, and recapitulation of themes. In the genre of the concerto, further, a third factor becomes operative and coordinated with the other two: that of opposition between the tutti and solo sections of a movement, opposition between the two “voices” conveying the musical discourse. Finally, yet one more factor fundamentally contributes to the creation of the classic instrumental forms, a factor commonly overlooked today but central in late eighteenth-century thinking, namely, the use of cadences to punctuate musical discourse. I do not claim that classic form is the product of punctuation alone but, rather, that it results from a complex interaction of all four factors—punctuation, voice, key, and theme. Punctuation, however, plays the central role in this interaction and provides an indispensable framework for understanding the roles played by the other three factors.

Cadence is a device to close a phase or the whole of a musical discourse. It symbolizes closure—not as an arbitrary signal but as the experience of reaching a goal. A stylistically competent listener, even if not musically literate (p.181) and unable to recognize and name the cadence, feels closure. Diverse musical factors—harmonic, melodic and contrapuntal, and metric—collaborate to produce this experience.

Because the function of cadences is to provide a sense of closure, they are naturally considered among the factors articulating musical form. But cadences can serve also to integrate form. The sense of an ending, to borrow Frank Kermode's phrase,4 is the single most important moment transforming a temporal succession of unrelated elements into a unified configuration of related phases. The ending is the essential function within the temporal form because with it the form gets its “point,” its goal. All the implications of earlier phases are finally explicitly realized, and thus the phases, whose relationship to one another might until then have been unclear, are now integrated into the whole.

The sense of closure provided by cadences is a matter of degree: cadences can be stronger or weaker. Experience shows that the stronger the cadence, the longer the time span for which it may provide closure. This remarkable fact has far-reaching formal implications. A musical style that develops a hierarchy of cadences, in which the relative strength of cadences is clearly differentiated and graded, will acquire a capacity to organize its phases hierarchically. The strongest cadences will close the discourse as a whole; the weaker ones will close parts of the discourse; still weaker ones will close parts of the parts. This sort of hierarchical organization of phases is the hallmark of the classical style.

There is a venerable tradition of regarding cadences as the single most important factor in the creation of musical forms. It extends from the beginning of the more systematic thinking about musical form in the late sixteenth century (to say nothing of even more ancient precedents in chant theory) to the very end of the eighteenth century, and it underlies what Carl Dahlhaus aptly named the rhetorical concept of musical form.5 Heinrich Christoph Koch, by far the most penetrating and representative theorist of musical form in Mozart's time, is still operating squarely within this tradi-tion.6 A modern reader of his three-volume Versuch einer Anleitung zur Composition, which appeared between 1782 and 1793, must be struck by how little interest Koch shows in such matters as a composition's tonal plan or thematic argument. Instead, Koch is guided by the rhetorical idea of musical composition, the vision of music as a discourse articulated like speech by means of stronger and weaker cadences into a hierarchy of parts:

Certain more or less noticeable resting points are generally necessary in speech and thus also in the products of those fine arts which attain (p.182) their goal through speech, namely poetry and rhetoric, if the subject that they present is to be comprehensible. Such resting points are just as necessary in melody if it is to affect our feelings …. By means of these more or less noticeable resting points, the products of these fine arts can be broken up into larger and smaller sections. Speech, for example, breaks down into various sentences through the most noticeable of these resting points; through the less noticeable the sentence, in turn, breaks down into separate clauses and parts of speech. Just as in speech, the melody of a composition can be broken up into periods by means of analogous resting points, and these, again, into single phrases and melodic segments.7

Koch's thinking about musical form rests on the central notion of punctuation. Form, for him, is “punctuation form.”

During the nineteenth century punctuation gradually lost its fundamental importance for the theory of musical form. Even Hugo Riemann's profound interest in it late in the century (interest that did not escape Nietzsche's attention)8 could not stop this decline: the change of paradigms that guided the thinking about music after 1800 was irreversible.9 Musicians stopped taking much notice of punctuation as the rhetorical concept of form was replaced by the notion of thematically and tonally driven musical logic—a change precipitated by the rapid disappearance of rhetoric from the intellectual horizon of educated Europeans and its replacement with philosophical aesthetics as the foundation of all theory and speculation about the arts. Once convention became suspect and romantics began to praise uniqueness above all else, it was inevitable that cadence, regarded as the most conventional of all musical phenomena, would give way to theme, considered the seat of compositional originality and individuality.

In thinking about classical form it is prudent, however, to recover what romanticism has obscured, and to consider punctuation seriously. This is not merely an act of courtesy toward our musical ancestors, paying attention to what they thought they were doing before imposing our own interpretations on their works. After all, there is no reason to believe that they were in a privileged position and understood their actions better than we could. We certainly have no such privileged access to our self-understanding; others may, and often in fact do, understand our actions better that we understand them ourselves. No, giving serious consideration to the rhetorical concept of form is simple economy of interpretative efforts: we should review the agent's own interpretation of what he is doing before expending efforts on our own interpretations; his just might turn out to be adequate, that is to say, more or differently illuminating than anything we can come up with ourselves.10

(p.183) The piano concerto was, of course, not the only instrumental genre to which Mozart gave his best efforts. Surely the G-Minor piano quartet, the six string quartets dedicated to Haydn, the last four string quintets, and the last four symphonies are among his most significant creations. In no other genre, however, did he leave as many masterpieces as in that of the piano concerto: K271 (January 1777) was arguably his first masterpiece in any genre and, together with the 1779 Sinfonia concertante, unquestionably the most impressive work of his Salzburg years. Among the series of twelve concertos composed for the four successive Viennese concert seasons between late winter of 1784 and early winter of 1786, at least nine works, K. 453 to K. 503, must be considered masterworks. Moreover, given the close proximity of instrumental concerto and operatic aria—in both a soloist is cast against an orchestral framework—it is appropriate that the composer for whom opera was central would choose the concerto when presenting himself to the public as a performer and would lavish on it some of his best inspirations.

Mozart at Play

Harmonic, melodic-contrapuntal, and metric factors all go into making a cadence. An adequate description of cadences in a given repertory must consider the harmonic functions of the several chords of which they are composed, the melodic content of the structurally important voices (in the eighteenth century that was soprano and bass), and the metric placement of the chords within measures. Here it will be useful briefly to summarize the results of an earlier investigation, in which I applied the late eighteenth-century notion of punctuation form to the allegros of Mozart's piano concertos in great detail.11

Mozart wrote two kinds of cadence: full cadences (ending on the tonic triad) and half cadences (ending on the dominant chord). (Occasionally a full cadence is transformed into a deceptive one when the final tonic triad is replaced by an unexpected harmony.) In a fast movement the last chord of a Mozart cadence is always placed on the strongest—the first—beat of the measure. The penultimate harmony in the full cadence is, of course, always the dominant, either plain or, in the great majority of cases, embellished with a 6/4 suspension; preceding the dominant in the full cadence is either one of the subdominant-function chords, or one of the tonic-function chords. In a half cadence the chords that may precede the dominant are the same as the ones that precede the dominant in a full cadence. The final tonic and the penultimate dominant of the full cadence, as well as the final dominant of the half cadence, are always in root position. Finally, the last (p.184) chord of a full cadence always has the prime not only in the bass but also in the soprano.

If Mozart was able to articulate highly complex hierarchical forms using only two cadences, it is because he knew how to calibrate the strength of a cadence. Recall that the stronger the cadence, the longer the time span for which it may provide closure. Both of the two basic cadences can be made stronger or weaker depending on the span that required closure. Strengthening is achieved by extending the duration of individual cadential chords. In the basic cadence each chord takes at least one beat of a quarter-note value, with the final chord taking an entire measure. All of these chords can be extended at will: extending the final tonic emphasizes the point of arrival; extending the dominant, whether penultimate or final, or the preceding chord, intensifies the expectation of arrival.

Extension of the final chord takes the form of an appendix. Extending any part of the cadence preceding the final chord results in internal extension or parenthetical insertion: either by extending the basic chords beyond their one-beat minimum or by inserting additional harmonies. The most impressive example of internal extension is, of course, that of the dominant-function chord with 6/4 suspension in the soloist's cadenza.

Just as extending the duration of chords strengthens the cadence, so contracting duration will weaken it. Contraction serves to connect more closely the phrase that is ending with the phrase that follows, so as to weaken the articulation between them. Since it is difficult to shorten the one quarter-note beat required for all but the final chord, contraction may affect only the final chord, which normally takes a whole measure.

In Mozart's music, contraction takes two forms: the link and elision. Two factors operate in normal articulation between two phrases. First, the final harmony of the first phrase falls on the first beat of the phrase's last measure, and the first melodic note of the second phrase falls on the first beat of the next measure, which is the first measure of the second phrase. Second, there is at least one beat and at most three beats of general rest in both the melody and the accompaniment at the end of the first phrase. A link consists of preserving the first factor and obliterating the second; this weakens the cadence. An elision weakens the cadence still further by obliterating both factors: the last harmony of the first phrase and the first melodic note of the second coincide on the same beat.

With two types of basic cadence (full and half), as well as various forms of extension (appendix, internal extension, and parenthesis) and contraction (link and elision) at his disposal, Mozart was able to project a discernible hierarchy of punctuation points and hence also a hierarchy of temporal (p.185) phases closed by these points. Knowing his punctuation tools, we can now examine and compare the musical forms he fashions with them to discover his generic norms, what David Rosen calls Mozart's “standard operating procedure.”12

In Koch's terminology, to be freely adapted here, the shortest independent temporal part of a composition—the part ending with the weakest, contracted cadence—is called a “phrase”; a series of phrases ending with a stronger kind of full cadence (either uncontracted, or extended, or both) is called a “period.” Mozart's basic concerto allegro form contains three periods of this sort, here referred to as the first, second, and fourth periods; the first normally corresponds to the first tutti, the second to the first solo, and the fourth to the vestigial third tutti and third solo together. Each of these three periods has an identical internal form. The strongest cadence is, of course, a full cadence at the end. The second strongest cadence in each of the three periods is a half cadence with an appendix. This divides the period into two sections. Each of these sections consists of two phrases, the first of which is closed by the weakest—because often unextended and always contracted—full cadence. The function of the first phrase is to begin the period. That of the second is to divide the period into two halves, that is, to close its first section. At the same time, since the second phrase ends with a half cadence and not with a full one, its closure is felt to be incomplete and thus prepares the arrival of the second section. The third phrase marks the beginning of this second section, and the fourth phrase ends the period.

The second and fourth tuttis, unlike the first one, do not constitute complete independent periods. They consist, rather, of extra closing phrases joined to the preceding periods. The cadences ending these tuttis are invariably stronger than those of the preceding periods. In the basic form the second tutti consists of only one phrase, and the fourth tutti consists of two phrases, the first of which closes with a cadence contracted but also tremendously extended by the soloist's cadenza. These additional closing phrases have a clear function: the second tutti divides the composition into two halves by providing an ending for its first part; the fourth tutti, which is longer than the second and contains in the cadenza the largest cadential extension of the movement, ends both the second part and the composition as a whole. Without the second tutti the movement would not have two parts; without the fourth the two parts would not be symmetrical and the movement would not have an appropriate ending.

The soloist's cadenza provides the climax of the movement, the moment when, compounding the high tension of a long-suspended dominant 6/4 with the excitement of improvisatory acrobatics, the solo performer reviews (p.186) the main events of the discourse now coming to an end. At the same time the cadenza ensures that the movement arrives at what has been the goal of all preceding events. The fact that the cadenza is improvised (whether actually or fictionally, no matter) compounds its climactic, revelatory character: until now the pianist, like an actor in a drama, played a part written for him by the composer; now he steps out of the prescribed role and speaks in his own name. For this one moment, the distance between player and personage being played disappears. Even when Mozart played the concertos himself (and he wrote many with himself in mind as soloist—in his Viennese maturity they were his main vehicle of self-presentation to the public as a player, and they contained his own greatest parts), most of the time he was playing the part of “Mozart,” the way Molière might be playing the part of Orgon, Alceste, Harpagon, M. Jourdain, or Argan.13 Only in the cadenza could “Mozart” take off the mask and reveal himself as Mozart. When the concertos are played by someone other than Mozart, as they inevitably must be, this point can be made even more clearly—which is why the wish to avoid stylistic incongruity on the part of those who improvise or write cadenzas seems to me profoundly misguided. What can be more thrilling than to hear Beethoven's personality unmistakably detach itself from the role the moment his cadenza for K. 466/i begins?

One phase of the concerto allegro remains to be considered: the second solo. For Koch this solo constituted a period in its own right, since its temporal dimension makes it comparable to the other three periods. But in most respects this is a highly irregular period. Unlike the other three, it may consist of as little as a single long phrase closing with a half cadence and an uncontracted appendix.

The basic concerto allegro form might be schematically represented as A1-A2-B-A3 (Table 1). The most striking feature of the form is the preoccupation with symmetry and balance that guides the division of the whole into two halves on every level. The composition consists of two parts, each of which contains two periods and an additional closing phrase or phrases. With the exception of the third, all periods consist of two sections, and each of these, in turn, of two phrases. The third period's irregularity strengthens rather than weakens the overall balance and symmetry, for by disrupting the order established in the first half of the composition, its restoration in the last period is all the more desirable and satisfying. Order that is never threatened can hardly be thematized and perceived. One half answering and balancing the other half, order imperiled but ultimately prevailing—this is what Mozart's sense of form requires.

One half answering and balancing the other, like two wings of a symmetrical (p.187)

TABLE 1. The basic concerto allegro punctuation form

Part I

Period 1 (regular)

Period 2 (regular)

Closing phrase:

  • ends with the uncontracted full cadence or appendix;

  • ends Part I and divides the whole into two parts.

Part II

Period 3 (irregular): ends with the half cadence and uncontracted appendix.

Period 4 (regular)

Closing phrase 1:

  • ends with the full cadence contracted and enormously extended by the cadenza.

Closing phrase 2:

  • ends with the uncontracted full cadence or appendix;

  • both closing phrases 1 and 2 end the whole.

facade: this kind of form may safely be called “architectonic,” as Jacques Handschin has done, provided we are not seduced by the metaphor into thinking that it is literally like architecture—that is, atemporal.14 On the contrary, answering and balancing, establishing order, disrupting and then restoring it—all of this crucially depends on the temporal disposition of events, on linear shaping of time, and on orientation toward a final goal, with temporary, hierarchically subordinate goals met and passed along the way.

It is important to note that the concerto allegro form turns out to be a large-scale variant of the most pervasive pattern of the era, one Leonard G. Ratner has called the “two-reprise” form.15 This pattern underlies much of the music written by Mozart and his contemporaries and is frequently found both on the level of a single phrase and of a whole movement. Central to this pattern is the ideal of bipartite, symmetrical balance, in which the second half answers the first and provides a resolution to its tensions and problems—it is, in short, an antecedent-consequent phrase writ large. Since a four-phrase period is the fundamental building block of the form, the pattern's most basic, minimal variant must consist of two such periods; further, the harmonic and thematic content of these two periods must be sufficiently similar to provide a sense of correspondence, and sufficiently different to justify the sense that the second does not merely repeat but answers and resolves the first. The simplest guise of this pattern fills the corresponding (p.188)

TABLE 2. Distribution of ritornellos and solos in the concerto allegro

Part I

Period 1

= Ritornello 1

Period 2

= Solo 1

Closing phrase

= Ritornello 2

Part II

Period 3

= Solo 2

Period 4

= (Ritornello 3), Solo 3

Closing phrases

= Ritornello 4 (with embedded cadenza)

phrases of the two periods with corresponding thematic contents, moves away from the home key for the second half of the first period, and resolves the resulting harmonic tension in a second period that remains in the home key from beginning to end. A “sonata allegro without a development,” as this is commonly called (where the first period is an “exposition,” the second a “recapitulation”), can be expanded in various ways. Adding a “development” (our third period) between the two periods might seem to threaten the bipartite ideal but actually does not do so when the first period is repeated and the last one is not, as would be customary. The resulting “sonata allegro” has an A1—A1—B—A2 pattern that is very close to the one we have identified in the concerto allegro.

The concerto allegro results from crossing sonata allegro form with such inherited generic requirements of the concerto as the alternation of framing tutti ritornellos and solo episodes, whereby the resulting form is guided by the ideal of bipartite balance (Table 2). The transformation of a sonata into a concerto allegro could not be more logical. When the soloist is given the main substance of a movement, it follows that the first period of the sonata, the “exposition,” becomes the first solo (period 2 in Table 1); the “development” then becomes the second solo (period 3), and the last period, the “recapitulation,” becomes the third solo (period 4). Since in a concerto the tonally stable (nonmodulating) ritornellos traditionally provide a frame for the frequently unstable solos, the three solos of the concerto require framing by four ritornellos. Instead of repeating the first period of the sonata, the composer casts the opening first ritornello in the form of a full regular period (period 1); in obedience to the generic tradition in which ritornellos do not modulate, the structural modulation away from home key is reserved (p.189) for the first solo only. By this point the composer will have used all of the components of sonata allegro form. But in a concerto the bipartite division is additionally clarified by a closing phrase at the end of Part 1 and two closing phrases at the end of the whole. It makes perfect sense that these are projected as ritornellos too: the second ritornello at the end of Part 1 and the final ritornello at the end of Part 2. There is, strictly speaking, no room for more than these three ritornellos (one each at the beginning, end of Part 1, and end of Part 2), no unoccupied phrase with a useful function to perform left. Therefore, the third ritornello, the one separating the second and third solos, survives in a vestigial form only: it does not get a separate phrase of its own but is usually marked briefly at the very beginning of the fourth period.

Order imperiled but ultimately prevailing: it is this sense of form that allows us in large measure to account for Karl Barth's profound observation that whereas “darkness, chaos, death, and hell do appear [in Mozart's music] … not for a moment are they allowed to prevail.”16 Barth continues his meditation:

The Mozartean “center” is not like that of the great theologian Schleiermacher—a matter of balance, neutrality, and finally, indifference. What occurs in Mozart is rather a glorious upsetting of the balance, a turning in which the light rises and the shadows fall, though without disappearing, in which joy overtakes sorrow without extinguishing it, in which the Yea rings louder than the ever-present Nay …. We will never hear in Mozart an equilibrium of forces and a consequent uncertainty and doubt …. This feature is enough to mark Mozart's church music as truly sacred. … Mozart … always achieved this consoling turn.

In large measure responsible for both the “upsetting of the balance” and the final “consoling turn” is Mozart's sense of linear temporal form and temporal disposition: initial balance is the prerequisite for the upset of balance, and that must be followed by the consoling turn. I am tempted to extend Barth's meditation and observe that Mozart's fundamental optimism, his sense that, though shadows will not be forgotten, things will turn out all right in the end, is not necessarily or exclusively religious. It is, rather, a sense of life characteristic of pre-revolutionary—pre-Terror—Enlightenment (Kant's view of history comes to mind as another example of the same trust, which similarly escapes the charge of naïveté because it is tempered by the same awareness of the possibility of tragedy). This Lebensgefühl has been utterly absent from art for so long now that we cannot but hear in Mozart's music a voice from a very distant and alien past.

(p.190) Yet Barth was right to suspect that the sense of life finding its embodiment in musical form had a religious dimension too. In a famous letter of 4 April 1787, addressed to his gravely ill father (Leopold died on 28 May), Mozart wrote:

As death, when we come to consider it closely, is the true goal of our existence, I have formed during the last few years such close relations with this best and truest friend of mankind, that his image is not only no longer terrifying to me, but is indeed very soothing and consoling! And I thank my God for graciously granting me the opportunity (you know what I mean) of learning that death is the key which unlocks the door to our true happiness. I never lie down at night without reflecting that—young as I am—I may not live to see another day. Yet no one of all my acquaintances could say that in company I am morose or disgruntled. For this blessing I daily thank my Creator.17

Too urbane and civilized to be morose or disgruntled in company, the author of this letter was yet no stranger to life's shadows. And while he wrote to offer his father encouragement, there is no reason to doubt the sincerity of a faith that allowed him to consider death to be the true goal of our existence and the final consoling turn. Nor is there any reason to believe that this faith was formed only during Mozart's last years. Our most profound convictions are formed and reformed over an entire lifetime. The same conviction of 1787 is present, albeit inchoate and naïvely expressed, in a letter the fourteen-year-old wrote from Bologna to his mother in Salzburg on 29 September 1770:

I am sincerely sorry to hear of the long illness which poor Jungfrau Martha has to bear with patience, and I hope that with God's help she will recover. But, if she does not, we must not be unduly distressed, for God's will is always best and He certainly knows best whether it is better for us to be in this world or in the next. She should console herself, however, with the thought that after the rain she may enjoy the sunshine.18

Like the cadence, the concerto allergro form may appear in its basic, extended, or contracted guise. The generic model reconstructed thus far was the normative blueprint underlying actual compositions. In considering a composition as a whole, it is important to keep in mind the functions performed by each of the essential blueprint phrases. Strictly speaking, each of these phrases is no more than a function slot, a place to be filled with one or more phrases capable of performing the given function. The extension of the basic form occurs when a slot is filled with more than one phrase. A (p.191) contraction occurs when a slot is left empty or when two successive slots are joined into one.

Examples of contraction are rare. They are most likely to occur in the first period, since the formal gaps and ambiguities introduced there can be filled in and clarified in the second and fourth periods. Unlike contraction, extension of the form—function slots being filled with more than one phrase each—is fairly common. Given that all phrases in any one slot should be capable of performing the functions of that slot, it is not surprising that they all close with the same kind of cadence; thus all phrases in the second-phrase position in the regular period end with a half cadence, whereas all the phrases in the other positions end with a full cadence. All internal phrases in a singlephrase position close with cadences that have been contracted to increase the sense of continuity and prevent too many interruptions to the flow of the discourse.

I see Mozart's form not as a rigid mold but as a flexible recipe with a few indispensable ingredients and procedures that a creative cook can supplement in a variety of ways; better yet (better, since this comparison captures the essential linear temporality of the form), I see it as a commedia dellʼarte scenario, which prescribes a few indispensable events and their order but leaves the artists at liberty to flesh out the plot with optional additional incidents in rule-governed improvisation. This vision of form—flexible recipe for making music rather than rigid mold to be filled with it—allows us to understand how Mozart's concertos can belong so unquestionably to a recognizable genre and yet exhibit such a high degree of individuality not just in material but also in form. It also allows us to appreciate the closeness of composition and improvisation in Mozart's creative thinking, and the essential rule-governed freedom of both.

“I hear Mozart … at play,” says Karl Barth, and he continues: “Beautiful playing presupposes an intuitive, childlike awareness of the essence or center—as also the beginning and the end—of all things. It is from this center, from this beginning and end, that I hear Mozart create his music. I can hear those boundaries which he imposed upon himself because it was precisely this discipline that gave him joy.”19

Mozart at Play

The harmonic content with which Mozart filled punctuation form, in concertos as in most other genres, is so predictable and schematic as hardly to warrant comment: in the first solo (or in the exposition of a sonata), the second phrase initiates the structural modulation away from the home key and (p.192) prepares the arrival of the second key by means of a secondary dominant; the third phrase confirms the arrival of the second key. In the second solo (or the development)—harmonically most unstable and variable—further modulations eventually issue in a retransition back to, and a dominant preparation of, the home key, the arrival of which is confirmed at the beginning of the third ritornello and third solo (or the recapitulation).

These basic facts are supremely important for the way music is experienced: the second half of a composition is felt to be more than a mere repetition of the first half because it is experienced as a resolution of the disturbance introduced by structural modulation. But while undeniably important, these harmonic events are too schematic and predictable to be of much interest. The individual, unique character of a composition resides in the melodic (thematic and motivic) content that fills its phrases. It is on this content that composer and listener alike primarily concentrate.

This is particularly true of a concerto allegro. Here a unique feature of the first half is that the two periods are not identical: the first solo, which modulates, cannot be a mere repetition of the first ritornello, which does not modulate. The thematic and motivic content of the corresponding phrases of these two periods can be similar, but they usually differ somewhat. These differences pose a problem in the second half of the form: since it, unlike the first half, contains only one regular period, all the melodic materials introduced in the first two periods somehow have to be accommodated and ordered in the final solo. The opening ritornello and solo provoke the question: How can these diverse materials be synthesized in a single period? The final solo answers the question. It is this question and answer that draw our attention. Here, again, the temporal order of events is essential. The composer's task is not simply to accommodate all the materials of the first two periods in the final one. It is, rather, to find a suitable, convincing, and logical temporal order in which to present the materials.

The general principle governing this temporal logic relates to the presence of function slots. The job of the first and third phrases is to provide a suitable beginning for, respectively, the whole period and its second half—to establish the subject and key of the discourse to follow. An antecedent-consequent phrase does this job particularly well, and more often than not it is a phrase of this sort that fills these two function slots. The job of the second phrase is to divide the period into two halves and prepare the arrival of the second half. An antecedent-consequent phrase would be inappropriate here. Something more flexible and capable of motivic fragmentation and tonal modulation is needed to get from one subject and key to the next. Finally, the job of the fourth phrase is to provide a suitable ending for the (p.193) whole. At this point the composition no longer needs a new subject of discourse; the appropriate phrase here, whether of the antecedent-consequent variety or not, need not be individually profiled but should contain emphatic cadential figures. Thus phrases of the first two periods, when recapitulated and synthesized in the last period, will reappear in the same function slot in which they had originally been introduced. When the third phrases of the first two periods get different thematic contents, for example, both themes will reappear in the third-phrase position in the last period. This rule has exceptions. But we need to understand such rules of the game before we can observe Mozart at play. This is always a riveting experience, whether he follows the rules or breaks them.

Among Mozart's instrumental compositions, his concertos stand out for their profusion of melodic invention. If the austere point of the fugue was to derive the whole discourse from a single idea, the ethos of the concerto is more relaxed and promiscuous: the more ideas, the merrier. The main focus of interest in the concerto allegros is the order in which melodic ideas are presented within each of the story's three tellings—the opening ritornello, the exposition, and the recapitulation—and playing with the order makes each of the tellings progressively more complete and logical. To demonstrate the full range of inventive solutions Mozart finds in playing this game would explode the limits of this book, but I can suggest something of the range by concentrating on a single radical movement from the piano concerto allegros.

Mozart at Play

No other movement departs as far from the generic norms as the Allegro of the Piano Concerto in C Minor, K. 491, dated 24 March 1786; hence none demonstrates more fully the extent of flexibility with which these may be treated.

The movement presents a unique case of formal extension: Mozart not only filled the individual function slots with several phrases but also repeated the entire sequence of slots. Table 3 shows the corresponding events of first ritornello, exposition, and recapitulation. Only the opening ritornello presents the regular sequence of four phrases or phrase groups. The first phrase of the exposition is missing, somewhat compensated for by a long, elaborate solo lead-in (mm. 100–118). By contrast, the third phrase-group (mm. 147–65) proceeds regularly to the fourth group (mm. 165–200), after which the sequence is repeated: instead of the expected second ritornello, the third phrase returns (mm. 201–20), followed by the return of the fourth phrase (mm. 220–65); only then, at m. 265, does the ritornello appear. In (p.194)

TABLE 3. Mozart, Piano Concerto in C Minor, K. 491: melodic content of the Allegro

First ritornello













( )

( )





Mozart at Play
















( )








Mozart at Play


















( )




( )







( )

NOTE: In each of the three systems, the first line indicates the measure with the downbeat on which the event marked in the second line begins. In the second line, capital letters mark significant or highly profiled melodic content—i.e., themes; lowercase letters are reserved for more conventional (transitional or cadential) motifs. Empty parentheses indicate appendices or prefixes whose content is either the same as that of the phrase to which they are attached or conventional. Numbers refer to the four phrases of the periods and the fifth (closing) phrase.

(p.195) the recapitulation, the first phrase is again lacking, this time without any compensatory lead-ins, and the double sequence of phrases three and four is further complicated—or rather, clarified—because this time the listener does not expect the ritornello earlier than when it in fact comes. The third phrase-group (mm. 391–428) is followed by what begins like the fourth phrase but then transforms in midstream into another second phrase (mm. 428–35); with its appendix (mm. 435–44) because, instead of closing with a full cadence, it gets, like any second phrase—and this is true only of second phrases—a half cadence (m. 435). This is followed by another third phrase (mm. 444–63) and then another fourth phrase (mm. 463–73). Immediately striking about both the exposition and the recapitulation is the drastic abbreviation of the first half (because of the missing first phrase) and the even more drastic amplification of the second half (the sequence of third and fourth phrases occurring not once, but twice). Even without the following ritornellos (the fifth phrase groups) the two “halves” are quite disproportionate.

It is not that the main subject is missing from exposition and recapitulation. Rather, since the main subject appeared twice at the beginning of the opening ritornello—first as the first phrase with a full cadence (mm. 1–13), and then as the second phrase with a half cadence (mm. 13–28)—it is possible to bring it back at the beginnings of the exposition (mm. 118–35) and recapitulation (mm. 362–81) as the second phrase only.

The main subject is also much in evidence in the final section of each of the three ritornellos. A variant of the subject closes each ritornello (mm. 91–99, 265–82, and 501–509, respectively; the haunting appendix-coda that closes the movement, mm. 509–23, is also based on the subject).20 Moreover, the subject itself, not a variant, opens the fourth phrase of the first ritornello (mm. 63–73) and the fourth phrase of the first solo (mm. 220–41), as well as being reserved in the last ritornello for the pre-cadenza phrase (mm. 473–86). In other words, the final ritornello recapitulates the last phrase-group of the opening ritornello, whereas the second ritornello limits itself only to the final phrase of the other two. Evidently, however, Mozart also wants the main subject toward the end of the exposition; hence its insertion at the beginning of the final phrase of the first solo (mm. 220–41). There it appears to be out of order, as the opening ritornello has led the listener to expect the main subject after, not before, the runs that close the first solo (that is, at m. 265, not 220). This lapse is corrected in the recapitulation, where it appears not before, but after, the soloist's closing runs (that is, at m. 473, not 463).

The larger point here, however, is that the main subject returns toward the end of the three ritornellos, given that it had been dropped in its firstphrase (p.196) form from the beginnings of the exposition and recapitulation. It is as if Mozart wanted to shift the subject from the beginning, its usual position, closer to the end. This, too, emphasizes the disproportion between the two halves of the story in each of the three tellings.

This shift of balance in the story may be related to the structure of the main subject itself (Example 4). The subject lacks the usual antecedent-consequent balance and instead accumulates chromatic tension in its drive toward the cadence. The emphasized melodic pitches—c′ (m. 1), f ♯′ (m. 4), e ♮′ (m. 6), d′ (m. 8), e ♭′ (m. 9), f′, and g′ (m. 12), c″ (m. 13)—suggest an ascending C-minor triad, initially destabilized by extreme chromaticism but driving inexorably toward the final c″. This drive toward the end is also underscored in the later part of the phrase by the first oboe, in counterpoint, chromatically descending from a ♭″ to c″. The final c″ thus becomes the goal for both lines. In its lack of balance and its drive toward the final note the subject thus both prepares and justifies the structural character of the entire movement—its forward drive and emphasis on endings rather than beginnings.

Between the beginnings and endings of the three tellings of the story the movement brings three second subjects (the only such instance in Mozart's music). One second subject appears in the opening ritornello (mm. 44–63),21 and two more in the exposition (mm. 147–65 and 201–20), each followed by its own closing phrase (mm. 165–200 and 220–65). How the three are to be integrated into the last telling of the story is a question Mozart raises in the first half of the movement. The answer is probably not the one most listeners would expect: in the recapitulation Mozart reverses the order in which he had originally introduced the three phrases.22 Otherwise, the logic of the recapitulation is impeccable: the two subjects of the first solo are recapitulated side by side and followed by the solo's first closing phrase; thereupon the subject of the opening ritornello is recapitulated separately and followed by the solo's second closing phrase, much abbreviated.

What centrally matters in this movement is not only that Mozart abbreviates the first half of the exposition and recapitulation and expands the second half. More important, he also in effect shifts the principal subject's main appearance (the one closed with a full cadence) from early to later in each section, getting the proper order wrong the first time around and correcting it at the end of the movement. The profusion of second subjects further works to delay the main subject. What all of this adds up to is a structure as tremendously and deliberately lopsided as the main subject itself, featuring a beginning abandoned as quickly as possible and a headlong rush toward the end. The imbalance certainly goes a long way toward explaining (p.197)

Mozart at Play

EXAMPLE 4. Mozart, Piano Concerto in C Minor, K. 491, first movement, mm.1–13; with analytic sketch below

(p.198) the Allegro's tremendous forward-directed energy, an energy not even the coda can quite dissipate.

The main interest of a concerto allegro rests with the order in which its melodic ideas are presented: this is the variable surface that focuses the listener's attention, to which punctuation and harmony provide the underlying and stable background. The point of the music is to tell an amusing, moving, and coherent story—a story with a beginning, middle, and end—and to tell it not once, but three times, with each successive version clarifying and closing the gaps left in the preceding version. Mozart's instrumental forms may have their ancestry in the circular Baroque patters of the concerto and the binary dance, but the composer transforms these patterns in such a way as to unbend the temporal circle into an arrow. The recapitulation is no longer simply a return; it is the necessary outcome, the final closing of gaps and reconciling of differences. The fundamental premise of Mozart's playing is that the linear flow of time from past to future matters: we do not get the point of what we hear unless we pay attention to the temporal order of the ideas.


(1.) Adolf Nowak has traced the transition in eighteenth-century music theory from thinking primarily in terms of “invention” to thinking in terms of “disposition”; see his “Der Begriff ‘Musikalisches Denken’ in der Musiktheorie (p.376) der Aufklärung,” in Neue Musik und Tradition: Festschrift Rudolf Stephan, ed. J. Kuckertz, H. de la Motte Haber, C. M. Schmidt, and W. Seidel (Laaber: Laaber Verlag, 1990), 113–22.

(2.) I am borrowing my terms here from Paul Ricoeur, Time and Narrative, 3 vols. (Chicago: The University of Chicago Press, 1984–1988).

(3.) Among the more recent English-language discussions of classical instrumental forms I have found the following to be particularly useful: Edward T. Cone, Musical Form and Musical Performance (New York: W. W. Norton & Co., 1968); Charles Rosen, The Classical Style: Haydn, Mozart, Beethoven, expanded ed. (New York: W. W. Norton & Co., 1997); Leonard G. Ratner, Classic Music: Expression, Form, and Style (New York: Schirmer Books, 1980); Mark Evan Bonds, Wordless Rhetoric: Musical Form and the Metaphor of the Oration (Cambridge: Harvard University Press, 1991); James Webster, Haydn's “Farewell” Symphony and the Idea of Classical Style: Through-composition and Cyclic Integration in His Instrumental Music (Cambridge: Cambridge University Press, 1991); 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); James Webster, “Sonata Form,” in The New Grove Dictionary of Music and Musicians, 2nd ed. (2001), 23:687–701; James Hepokoski, “Back and Forth from Egmont: Beethoven, Mozart, and the Nonresolving Recapitulation,” 19th-Century Music 25 (2002): 127–54; and Hepokoski, “Beyond the Sonata Principle,” Journal of the American Musicological Society 55 (2002): 91–154.

(4.) Frank Kermode, The Sense of an Ending: Studies in the Theory of Fiction (Oxford: Oxford University Press, 1967).

(5.) Carl Dahlhaus, “Der rhetorische Formbegriff H. Chr. Kochs und die Theorie der Sonatenform,” Archiv für Musikwissenschaft 35 (1978): 155–76.

(6.) Heinrich Christoph Koch, Versuch einer Anleitung zur Composition, 3 vols. (Leipzig and Rudolstadt: Böhme, 1782–93).

(7.) Heinrich Christoph Koch, Introductory Essay on Composition: The Mechanical Rules of Melody, Sections 3 and 4, trans. Nancy Kovaleff Baker (New Haven: Yale University Press, 1983), 1.

(8.) Friedrich Nietzsche, Der Fall Wagner, 11.11–5, in Werke: Kritische Gesamtausgabe, vol. 6/3, ed. G. Colli and M. Montinari (Berlin: de Gruyter, 1969), 32.

(9.) We are indebted to Dahlhaus for a penetrating analysis of this change. See, in particular, his The Idea of Absolute Music (Chicago: The University of Chicago Press, 1989) and Nineteenth-Century Music (Berkeley: University of California Press, 1989).

(10.) Cf. Richard Rorty, “Texts and Lumps,” New Literary History 17 (1985): 1–16.

(11.) “The First-Movement Punctuation Form in Mozart's Piano Concertos,” in Mozart's Piano Concertos: Text, Context, Interpretation, ed. Neal Zaslaw (Ann Arbor: The University of Michigan Press, 1996), 239–59, as well as “Mozart's (p.377) Concerto Andante Punctuation Form,” Mozart-Jahrbuch 1998: 119–38, where I attempted to reconstruct how Mozart thinks in a slow movement.

(12.) David Rosen, “The Composer's ‘Standard Operating Procedure’ as Evidence of Intention: The Case of a Formal Quirk in Mozart's K. 595,” Journal of Musicology 5 (1987): 79–90.

(13.) For a different view of this relationship, see Peter Gülke, “Triumph der neuen Tonkunst”: Mozarts späte Sinfonien und ihr Umfeld (Kassel-Stuttgart: Bärenreiter-Metzler, 1998), 54–68.

(14.) Jacques Handschin, Musikgeschichte im Überblick (Lucerne, 1948).

(15.) Leonard G. Ratner, Classic Music: Expression, Form, Style (New York: Schirmer Books, 1980).

(16.) This quotation and the one following are from Karl Barth, Wolfgang Amadeus Mozart (Grand Rapids: W. B. Eerdmans, 1986), 53 and 55f.

(17.) Emily Anderson, ed. and trans., The Letters of Mozart and His Family, 3rd ed. (London: Macmillan, 1985), 907.

(18.) Anderson, ed. and trans., The Letters of Mozart and His Family, 163.

(19.) Barth, Wolfgang Amadeus Mozart, 16.

(20.) The autograph (London, Royal College of Music, Ms. 402) shows that Mozart himself considered the appendix a real coda. Instead of writing out mm. 492–508, Mozart made a sign to use mm. 82–98 there. Before m. 99 he drew a double barline, topped it with fermatas, and wrote above and below the score, “Coda.”

(21.) The autograph shows that this was something of an afterthought. Originally, Mozart went directly from m. 43 to the present m. 63, that is, composed no third phrase in the ritornello. But he changed his mind even before the ritornello was completed and wrote the second subject at the present m. 91.

(22.) Unexpected order reversals abound in Mozart piano concertos from early on. The obvious collorary here is the pleasure the composer took as a young man in jumbling proper word order in his letters and transforming perfectly correct but conventional sentences into slightly bizarre and absurd, though still recognizable, text. See, for instance, a letter of 21 August 1773, written from Vienna to his sister in Salzburg, which he signed “Gnagflow Trazom,” or the message of 16 January 1773, written to his sister from Milan: “I for have the primo a uomo motet compose which to tomorrow at Church the Theatine perfomed be will.” Anderson, ed. and trans., The Letters of Mozart and His Family, 239–40 and 226.