Archive for the ‘music analysis’ Category

Key-Related Idioms in Mozart’s Music: A Peek into his Creative Process?

June 19, 2011

Paper to be presented at the annual meeting of the American Society for Music Theory in Minneapolis, October 27–30, 2011, in cooperation with Dr. Saharon Rosset of the Department of Statistics and Opertaions Research, Tel-Aviv University.

Is the choice of key just a marginal aspect of musical composition or could it also significantly interact with musical substance? The issue has long intrigued music theorists, musicians and casual listeners. Whereas the traditional discipline of “key characteristics” examines the connections between specific keys and modes of expression, newer research has started to consider associations between keys and tangible musical structures. However, a systematic, data-driven and rigorous investigation of the correlations between key and structure throughout a composer’s body of works has never yet been attempted.

In this paper, we demonstrate the crucial role of key choice in determining concrete structural features of musical substance as exemplified by the compositions of Wolfgang Amadé Mozart. By performing an extensive survey of harmonic, melodic and rhetoric phenomena in Mozart’s works, we show that associations between keys and musical matter are represented practically at all levels of his compositional thinking, amounting to a statistically significant total.

According to our findings, key-relatedness in Mozart’s music tends to intensify with time. Additionally, we show that by tracking the behavior of “key-related idioms”, one may gain a novel insight into some intriguing chronological, aesthetical and semantic aspects of Mozart’s creative process.


Structural Deformation as a Token of Undercurrent Humor in Mozart’s Instrumental Rondos

June 19, 2011

Paper to be presented at the 7th European Music Analysis Conference (EUROMAC) in Rome, September 29 – October 2, 2011

“Non-referential” humor is prone to manifest in structural irregularities capable of upsetting the listener’s “preparatory set” of expectations (Meyer 1956). Applying this to the First Viennese School, what would typically come to mind are Haydn’s witty formal distortions as, e.g., in the finales of the ‘Joke’ Quartet and the ‘L’ours’ Symphony. In accordance with the widespread view that, among Mozart’s oeuvre, even instrumental works constitute “operas in disguise”, it may be surmised that Mozart’s humor is essentially operatic, opting for caricatures of characters and situations.

In reading Ein Musikalischer Spaß K. 522 as a “parody of compositional inadequacy”, Lister (1994) claims the work to have a virtual “hero”, the incompetent Kapellmeister it conjures up. By echoing topics derived from the style of opera buffa, instrumental pieces may incorporate comic elements without necessarily reenacting farcical stage situations. This has been demonstrated, e.g., with regard to subtle “slapstick effects” in the opening theme of the String Quintet K. 593 (Ratner 1980), to buffa-esque gestures in several of Mozart’s piano concertos (Allanbrook 1996) and to the buffo topic in the finale of the ‘Prague’ Symphony (Kenpler 1994, Sisman 1997).

However, it is also possible to trace in Mozart’s music a “non-referential” type of humor, which can be shown to operate along the lines of such complementary concepts as stylistic “markedness” (Hatten, 2004) and structural “deformation” (Hepokoski and Darcy, 2006). In Mozart’s instrumental rondos and sonata-rondos, deviations from formal ‘defaults’ are generally far more pronounced than in his sonata movements proper, and, in observing a selection of such irregularities, I endeavor to challenge the implicit common view that the composer’s musical wit necessarily drew on emulating operatic or opera-like procedures.

In the final rondo of the Trio Piano K.498 (‘Kegelstatt’), a melodic figure, originally “tucked away” inside one of the couplets (mm. 139–145), gradually assumes a higher degree of formal significance throughout the movement, thus provoking the image of an initially shy figure, becoming more and more obtrusive. Whereas this subtle manipulation, which may be likened to Meyer’s concept of “hierarchic migration” (1989), hardly seems to relate to any of the buffa-esque procedures normally utilized by Mozart, it nevertheless lends itself to a humorous interpretation in terms of a witty violation of thematic hierarchy within a piece.

Special attention will be given to the second movement of the Sonata for Piano and Violin K. 302. I will show that the formal ambiguity of measures 17–36, instead of being resolved, is constantly amplified throughout the movement, producing two concurrent, albeit incongruent, overall formal designs, somewhat reminiscent of a cubist painting.

Whereas Haydn’s sense of humor was considered already during his lifetime an essential prerequisite for an understanding of his music (Junker 1776), Mozart’s humor seems to have been less widely acknowledged as a central feature of his style. Allowing for Johann Georg Sulzer’s threefold distinction among ‘low’, ‘middle’ and ‘high’ comic (1777), K. 522 clearly falls under the former category, involving the ludicrous and farcical. Formal irregularities of the kind discussed in this paper, on the other hand, give rise to a more refined ‘middle’ (witty) humor, which may not necessarily be immediately perceivable as such.

Interrelation between elementary cognitive schemata and melodic surface in Mozart’s music

June 19, 2011

Paper to be presented at the annual meeting of the Israel Musicological Society in Raanana, July 3–4

A widespread view in recent research on 18th century music is that the melodic surface of a given composition results from combining and elaborating on a number of elementary cognitive schemata, assimilated by composers of the period during early stages of their career. A possible implication of this approach is that melodic events at the foreground level should be considered variable and open for modifications. This stance is taken up by a current approach in musical interpretation, promoting improvisation on the basis of the original text. My research into the correlation between key choice and musical structure in Mozart’s music supports a contrary view. Apparently, Mozart would perceive two melodic formulations as equivalent (and, consequently, tend to use them in connection with the same key) on the basis of literal resemblance at the surface level and not due to the implementation of common deep-level schemata. This, in turn, invites a closer examination of the concept of “decorability” and of the conditions under which melodic formulations in Mozart’s music permit or prohibit modifications.

Mozart and the Grief Process – Interpreting the String Quintet in G Minor K. 516

April 18, 2009

Mozart and the Grief Process – Interpreting the String Quintet in G Minor K. 516

Uri Rom



This article appears in the collection: Interpretare Mozart, Conveno Internationale di Studi, Milano, 19-21 Maggio 2006, ed. by Mariateresa Dellaborra, Guido Salvetti & Claudio Toscani, Libreria Musicale Italiana, Lucca 2007, pp. 381-387. See:

In this paper I will address questions of continuity and process building in Mozart’s string quintet K.516. I will try to show that this work, probably more than any other composition by Mozart, exemplifies the concept of an overall structure, crossing the borders between the individual movements. I will also try to show that, allowing for a metaphorical interpretation of music along the lines of Nelson Goodman’s symbol theory[1], the G minor string quintet may be viewed as a representation of grief process as described by modern psychology.

One feature which has been recognized in almost all analyses of Mozart’s string quintet is the quite unorthodox succession of movements, shown in the table below:

I Allegro

II Menuetto – Allegretto

III Adagio ma non troppo

IV Adagio/Allegro

This succession of movements shows two outstanding features, differing from the prevailing norm in Mozart’s oeuvre. The minuet appears here before, rather than after the slow movement; further, the finale is preceded by a slow introduction which, due to its extraordinary length and intensity, could be almost considered as a further slow movement. None of these procedures is in itself without precedent in Mozart’s works.[2] The combination of both, however, resulting in two consecutive slow movements, is unique to the G minor quintet. Whereas almost all discussions of Mozart’s G minor string quintet point out the unusual features of its overall structure, analyses have hitherto failed to observe that there is possibly a method to the apparent irregularity of the quintet.

Looking at the list of Tempi, we may note that it exemplifies a clear directionality from fastest to slowest. Beginning with the Allegro of the first movement and ending with the Adagio of the introduction to the finale an overall process of retardation spans the entire work, offering an absolute exception in Mozart’s Works and possibly also in the entire Viennese classic. The final rondo allegro is excluded from this process and in the course of the following discussion, I will also try to show why.

My claim is that this succession of Tempi could not be a chance product of Mozart’s compositional process: note the exactness with which Mozart differentiates between the Allegro of the first movement and the Allegretto of the minuet and – even finer – between the Adagio ma non troppo of the third movement and the Adagio of the slow introduction to the finale. It seems to me quite implausible that Mozart should have gone into introducing such irregularities in the order of movements and into making such subtle distinctions among the tempi, had he not consciously intended the gradual process of retardation in the quintet.

However, four tempo instructions are still not enough to create a musical process which is also accessible to the listener. In order for that to happen, the composer has to establish a sense of continuity in the work, reaching beyond the movement borders, so that each tempo may be heard in relation to the preceding as well as to the following tempo, in spite of the silences between the movements. Mozart masters this apparently impossible task by applying a variety of musical resources at the joints between the movements, as I would now like to demonstrate.

The first movement ends after a long passage in piano with two abrupt forte chords, giving an impression of interruption rather than conclusion. The second movement opens with a quite unusual dominant chord in the first inversion as an upbeat; this harmonically unstable beginning makes the movement sound as if starting from the middle.[3] Both the abrupt ending of the first movement and the open beginning of the second charge the silence between the two movements with tension: one hears the minuet somehow as a direct continuation of the first movement, also perceiving the new tempo, Allegretto, in relation to the preceding Allegro.

The minuet closes on the middle g in piano, while the Adagio ma non troppo starts on the same note in the same dynamics, only now embedded in an E-flat harmony. I hear also this transition as a kind of continuation, a bridge between the movements, making, once again, the new tempo sound as a calming down of the minuet tempo.

The third movement ends with an E-flat chord with a B-flat as upper tone; this disposition, quite unusual for a concluding chord, has something open about it, calling for a continuation. The next movement – the slow introduction to the finale – promptly responds with the very same chord, transposed a third higher to G minor. The repetitive semiquavers, accompanying the coda of the third movement, are answered by the quavers at the beginning of the slow introduction in an even slower tempo, bringing thus the overall retardation process of the quintet to its lowest point.

Also within the second movement there is a point which poses a challenge to the sense of continuity in the quintet: the transition from minuet to trio and back often involves a break in the continuity of sound, yielding two loosely connected sections. In the string quintet Mozart avoids this break by applying a simple but ingenious “trick”: the closing phrase of the minuet, transposed to G major, becomes the opening phrase and the thematic nucleus of the entire trio, thus making the trio sound not as an independent part but rather as a subordinate section of the minuet.

I have tried to show that the process of retardation in the G minor string quintet, obviously intended by Mozart, does not remain an abstract scheme but is sustained by a variety of subtle compositional measures, integrating the silences between the movements into an arch of musical tension, spanning the entire work. Moreover, the hypothesis of an overall process in the quintet supplies a plausible framework, within which many of its outstanding features suddenly fall into place: Not only the unusual succession of movements, but also the rather irregular harmonies at the beginning of the second as well as at the end of the third movement and the thematic connection between minuet and trio may be interpreted as means to establish the quintet’s global musical process.[4]

Having acknowledged that, we may ask ourselves what Mozart’s intentions behind the quintet’s global process might have been. With Mozart placing in his letters such an immense emphasis on the issue of expression and on music as means to convey ideas and feelings,[5] it is highly unlikely that he should embark on such an irregular musical conception without also wishing to express exceptional emotional contents.

According to my interpretation of the quintet, it is possible that Mozart intended to express in it emotions and mental processes connected with the experience of loss and grief. How much of this was actually consciously planned by Mozart, and to what extent his musical-emotional intuition may have been involved, will probably remain undeterminable. I certainly do not wish to suggest that the quintet should be regarded as an autobiographical work: whatever personal experience might be behind it, the quintet addresses emotions in a generalized way.

Psychologists generally agree that the grief process proceeds in phases, stressing, however, the fluid, non-linear as well as highly individualized character of the process. Bearing this in mind, the following phases may be sketched:[6]

Numbness (shock, denial, flight into hectic activity)
Searching and yearning (anger, pain, anxiety, ‘pangs’ of grief)
Depression (withdrawal, disorganization)
Recovery (acceptance, reorganization)

The first phase, numbness, is seen as a protective reaction, acting as a barrier to immediate pain. Some writers speak of this phase in terms of denial of the loss, mentioning hectic activity as a possible symptomatic reaction in order to block overwhelming emotions. The second phase, an inner search for the lost person, is characterized by strong, violent emotional eruptions of anger and anxiety. Once the intense pangs of grief subside, feelings of apathy and despair predominate. Disorganization as well as withdrawing from others are mentioned in connection with this phase. After having accepted the loss, reorganization of the self and of life may take place, leading to a recovery from the state of grief.

A generally acknowledged aspect of the grief process is the transition from the hectic, hyperactive reactions at the beginning of the process to the apathy of the depressive phase. In my opinion, this progression could be very well matched by the quintet’s overall process of retardation. Further, I would like to show that the attributes of the different movements of the quintet may also be viewed in a rough analogy to the different phases of grief.

Although the adjective “numb” could hardly apply to the first movement, it definitely exemplifies the kind of flight to hectic activity, mentioned in connection with the first grief phase. In my opinion, Mozart creates here a musical representation of obsession which is reflected in every aspect of the music. Most characteristic of the obsession in the first movement is surely the unyielding accompaniment of quaver repetitions, underlining over 70% of the movement and constituting a predominant element in it. Even passages which are harmonically and melodically rather relaxed retain through this accompaniment a tense, obsessive character. Considering that these quavers represent obsession, performers might be well advised not to play them too unobtrusively or lightly.

Another unusual feature of the first movement is its secondary theme which, very atypically, starts in the main tonality G minor, only subsequently modulating to the “correct” tonality of B-flat major. This is no usual secondary theme: it dominates the greater part of the development section as well as the coda. According to my view, this theme assumes the function of an idée fixe, haunting the listener throughout the movement. Mozart’s clinging to the tonality of G minor and shunning the modulation when it is due, may be also seen, in my opinion, as demonstrating an obsessive behaviour on a formal level.

The second movement, with its often interrupted phrases, harsh accents on the weak beats and abrupt changes of dynamics may represent the second phase of the grief process, the phase of violent “pangs” of grief, of yearning, anger and anxiety. The rather smooth G major trio offers a contrast to the character of the main minuet, but, as mentioned above, it draws thematically on the last phrase of the minuet and is therefore perceived as a subordinate section, a repose in the midst of the emotional tumult rather than an outcome.

The third movement, Adagio ma non troppo, is one of the very few movements among Mozart’s chamber music to be played throughout with mutes.[7] This gives the movement an aura of detachment and withdrawal from reality. After the first two movements, which are rather dense and at times almost breathless, the third movement strikes us as a succession of fragmented phrases, separated by frequent silences.

The difference to the customary adagio texture may be seen by comparing the third movement of K.516 with the slow movement of the “Hunt” quartet K.458 written about two and a half years earlier. In the Adagio from the quartet, Mozart “sews” adjacent phrases together using bridge elements; at the analog points in the quintet’s Adagio ma non troppo, Mozart leaves silences. The muted character of this movement, the frequent silences and the fragmented structure seem to correspond convincingly to the phase of disorganization and depression.

The slow introduction to the finale carries the depressive character of the third movement to a powerful emotional climax in bars 20–25 with the longest crescendo passage in the entire quintet. These bars seem to me to represent the painful moment of emotionally acknowledging the loss after having denied, objected to and grieved it; the moment in which the inner parting with the lost object eventually takes place.

Most puzzling is the fact that Mozart uses here, in the moment of culmination, the same harmonies as at the very beginning of the quintet, applying an identical chromatic sixths progression. This chromatic progression, originating from the standard lamento bass figure, here displaced into the upper voice and harmonized in an unorthodox manner, may be regarded as the harmonic “motto” of the entire quintet. Apart from the appearances in the first movement and in the slow introduction to the finale, it also recurs in bars 14–16 of the minuet in another tonality.

In addition to this harmonic “motto” Mozart creates a thematic unity in the string quintet by means of a melodic germ, recurring in all movements. Rather unspectacular, this motive, consisting of a segment of a descending diatonic scale beginning on the fifth or on the sixth scale tone, assumes in the quintet a central thematic role, appearing in each of the movements at conspicuous positions (first movement, measures 30–33, second movement, beginning, third movement, measures 18–21 and fourth movement, beginning).

The semantic role of this figure in Mozart’s works may be deduced from two occurrences in the Magic Flute, both remarkably enough also in G minor. The middle section of the first aria of the Queen of Night and Pamina’s despair aria No. 17 use the very same motive, connecting it with the subject of loss, despair and death.[8] The first movement and the slow introduction to the finale are further connected by a number of motives recurring in both. Some of these cross relations (as, for instance, between measures 18–19 of the first movement and 30–31 of the fourth) have not been pointed out so far.[9]

This thick network of harmonic and thematic cross-relations between the beginning of the quintet and the introduction to the finale may support an interpretation the quintet as an emphatically cyclic work. The overall process of retardation eventually leads us right back to the very beginning, thus supplying a striking musical-metaphorical representation of the fact that the cathartic confrontation with the loss at the end of the quintet is, in fact, the resolution of the same painful emotions, originating from the first movement.

Finally, I would like to point out that an interpretation of the string quintet along the lines of the grief process may also offer an explanation to an aspect of the quintet which deeply disturbed almost all analysts writing on the work. The “lieto fine”, the rather jolly and exuberant G major finale, seemed to many rather inadequate for a work starting off so gloomily, and quite a few attempts have been made to describe it as less jolly than it actually is in order to make it fit better into the quintet. Here may be mentioned Einstein’s “disconsolate major”, Hildesheimer’s “desperate happiness” as well as Bodenheimer’s demonizing interpretation, hearing this finale as a cynical, calculated parody.[10]

Considering these quite extreme interpretations, my hypothesis, based on an analogy to the grief process, seems to offer a much more straightforward view on the finale. Having finally overcome the confrontation with the loss in the slow introduction, a reorganization of life may now take place, this being convincingly depicted by the “back to normal” style of the finale. According to this view, not a failing sense of tragedy, but rather a genuine psychological understanding led Mozart to end one his most moving compositions with a bright, life-affirming rondo in G major.


1. Nelson Goodman’s views, expressed, among others, in Nelson Goodman, Catherine Z. Elgin, Reconceptions in Philosophy and other arts and sciences, Hackett, London 1988, have been elaborated upon by a number of writers, placing a special stress on metaphorical functions in music. A selection may include: Robert S. Hatten, “Metaphor ‘in’ Music”, in Musical signification: Essays in the semiotic theory and analysis, ed. by Eero Tarasti, Mouton de Gruyter, Berlin 1995, as well as Simone Marenholz, Musik und Erkenntnis, Metzler Stuttgart 1998, and the collection of essays Klang-Struktur-Metapher. Musikalische Analyse zwischen Phänomen und Begriff, ed. by Oliver Schwab-Felisch, Michael Polth and Christian Thorau, J.B. Metzler, Stuttgart 2000.
2. Placing the minuet before the slow movement is not unusual for the Viennese Classic. Mozart recurs to this procedure in half a dozen string quartets and possibly also in the C major quintet K.515, where the state of the autograph doesn’t permit an unequivocal determination of the succession of movements. A slow introduction to the finale occurs in a few serenades of the Salzburg period, consisting, however, of more than four movements, the slow introduction invariably following a minuet.
3. Whereas Mozart very often uses the fifth scale tone as an unaccompanied upbeat (as in the minuet of the “Linz” symphony K.425), he very seldom introduces a complete non-tonic chord at the beginning of a movement, another rare example being the minuet of the C major string quintet K.515.
4. Accepting this hypothesis has a rather trivial but yet crucial implication for the performance of the quintet: whereas interrupting the performance of any work for retuning the instruments may prove disadvantageous for the listener, I would strongly advise to refrain from such an interruption while performing the G minor quintet, as this will inevitably ruin what Mozart went to such pains to establish: the sense of continuity throughout the entire work as well as the overall process of retardation.
5. In a letter to Aloisia Weber from July 30th 1778, Mozart explains to her how to interpret his Andromeda scene Ah, lo previdi (K.272), stressing: “al più le raccomande l’espressione”. In an early letter to his father (November 8th 1777) Mozart claims he can achieve with sounds what a poet expresses with words, a dancer with pantomime and a painter with colours. An analysis of this letter appears in Georg Knepler Wolfgang Amadé Mozart, Annäherungen Henschel, Berlin 1991, pp. 28-30.
6. This presentation is based on views proposed in Colin Murray Parkes, Bereavement: studies of grief in adult life, Tavistock, London 1972, John Bowlby, Attachment and loss, Penguin, London 1980 and Verena Kast Trauern, Phasen und Chancen des psychischen Prozesses, Kreuz, Stuttgart 1986. A good overview of leading grief theories is presented in the Handbook of Bereavement ed. by Margaret S. Stroebe, Wolfgang Stroebe and Robert O. Hansson, Cambridge University Press, Cambridge/New York 1993.
7. The only further examples are the slow movements of the string quartets K.168 and K.171 as well as the string quintet K.174, all three works composed in 1773.
8. Analyzing the complete corpus of Mozart arias in G minor, Steven B. Jan points out that subjects of loss and separation are typical for the last five ones, composed during Mozart’s last decade. In my opinion, the preoccupation with those topics may also be transferred to Mozart’s instrumental compositions in G minor created during this period. This view may, in turn, confirm the interpretation of the G minor string quintet in connection with the grief process (see: Steven B. Jan, Aspects of Mozart’s Music in G Minor, Toward the Identification of Common Structural and Compositional Characteristics, Garland, New York 1995).
9. For a discussion of the thematic and rythmic cross references in the quintet see also: Marius Flothuis, Wolfgang Amadeus Mozart. Streichquintett g-Moll, KV 516 (=Meisterwerke der Musik 44), Wilhelm Fink Verlag, München 1987, pp. 23–24.
10. Alfred Einstein: Mozart, Sein Charakter, sein Werk (Frankfurt am Main 1967, p. 193); Wolfgang Hildesheimer, Mozart (Frankfurt am Main 1977, p. 178); Aron Ronald Bodenheimer: „Der trostlose Tröster“ (Über Mozart, edited by D. Klose, Stuttgart 1991, p. 316).

Mozart’s choice of key – a key to his creative process?

April 16, 2009

W. A. Mozart’s extremely accurate absolute pitch is a well established fact. Is it possible that this aspect of his hearing influenced the way Mozart conceived his music? Given Mozart’s  outstanding musical memory (which enabled him, for instance, to write down Allegri’s Miserere after just two hearings), one could proceed from the assumption that Mozart would ‘store’ his own music, as well as other music he had heard, at the pitch he had originally conceived/heard it, that is, in association with a specific key. When coming to compose a new piece in a given key, Mozart  would obviously draw on his memory to supply him with ideas. If Mozart’s mental ‘archive’ was somehow sorted by keys, it is natural to ask whether he would tend to use similar material in the same key over and over again.

This hypothesis was put forward by Wilhelm Gloede in 1993 in his article “Motivstruktur und Tonart bei Mozart” (“Motif structure and key in Mozart’s music”, Archiv für Musikwissenschaft 50/1, pp. 26–43). Gloede observes a correlation between five melodic structures in Mozart’s music and five corresponding keys. This correlation, however, is not conclusive; what Gloede’s data indicate could be best described as a fairly strong statistical tendency of certain motifs to appear in specific keys.

My forthcoming dissertation project on Mozart’s use of the keys investigates further instances of the phenomenon of key dependence in Mozart’s music. This site offers summaries of articles I have already published on the subject. The article “An ‘E-flat Major Motif’?” examines the various occurrences of a six note motif in Mozart’s works, showing it to be bound to the key of E-flat major. Further articles, dealing with various key-bound structural elements, are to follow.

A major question involved with the evaluation of Mozart’s use of keys is whether the wide variety of pitches and tuning systems during his lifetime may have interfered with the way he perceived and memorized music in a certain key. I will try to argue that – while pitch fluctuations undeniably modify the actual sonority of a key – Mozart must have developed a firm concept of keys, filtering much of the potential distraction caused by such fluctuations.

Furthermore, the entire complex of key characteristics should be taken into consideration in connection with the phenomena examined here. Following the reduction of the number of modes to just two (major and minor), an elaborate discipline emerged during the 17th and the 18th centuries, assigning different characters to many of the (structurally identical) major and minor scales. Does the concept of key characteristics supply an adequate framework for considering Mozart’s repeated employment of similar material in the same key? While one couldn’t refute that Mozart – like many of his contemporaries – might have been influenced by key characteristics, I will try to show that many types of correlation between key and concrete structural elements in his works are better described by the model presented above than by traditional key characteristics.

An intriguing question is whether one should expect concrete relations between key and musical structure to emerge in the works of just any composer in possession of absolute pitch and a strong musical memory, further, whether such relations should be deemed impossible in composers without these skills. More concretely, Mozart’s employment of key dependent structural elements should ultimately be considered within a broader perspective of the First Viennese School before the phenomenon could be evaluated properly. While my investigation is primarily concerned with Mozart’s music, some instructive evidence has been gathered from essays on Mozart’s contemporaries, suggesting that Mozart’s treatment of the keys may not have been qualitatively different from that of other composers.

Should a future, more extensive examination of key dependent elements in the works of other composers produce sufficient evidence, a reassessment of the phenomenon of key dependence in Mozart’s music might be due. For the time being, it seems that Mozart’s oeuvre is telling a story of its own, and that in this story the key in which a work is written plays a more concrete role in determining its musical structure than has hitherto been suspected.