A surviving letter of Mozart's to his father Leopold (31 July 1778) indicates that he considered composition to be an active process, the product of his intellect and carried out under conscious control:
In other words, a popular stereotype about creative artists, that they passively wait for "inspiration" to strike, is probably inapplicable to Mozart. For discussion of this stereotype as applied to Mozart, see below.
Mozart often wrote down sketches, ranging in size from small snippets to extensive drafts, for his compositions. Although many of these have not survived, having been destroyed by Mozart's widow Constanze, about 320 sketches and drafts are extant, covering about 10 percent of the composer's work.
Ulrich Konrad, an expert on the sketches describes a well-worked-out system of sketching that Mozart used, based on examination of the surviving documents. Typically the most "primitive" sketches are in casual handwriting, and give just snippets of music. More advanced sketches cover the most salient musical lines (the melody line, and often the bass), leaving other lines to be filled in later. The so-called "draft score" was one in an advanced enough state for Mozart to consider it complete, and therefore enter it (after 1784) into the personal catalog that he called Verzeichnüss aller meiner Werke ("Catalog of all my works"). However, the draft score did not include all of the notes: it remained to flesh out the internal voices, filling out the harmony. These were added to create the completed score, which appeared in a highly legible hand.
This procedure makes sense of another letter Mozart wrote to Leopold, discussing his work in Munich on the opera Idomeneo (30 December 1780), where Mozart makes a distinction between "composed" and "written":
In Konrad's view, Mozart had completed the "draft score" of the work, but still needed to produce the completed, final version.
Of the sketches that survive, none are for solo keyboard works. Konrad suggests that "improvisation [at which Mozart was highly skilled; see below] or the actual trying out of particularly challenging imaginative possibilities could compensate in these cases for the lack of sketches.
Mozart evidently needed a keyboard to work out his musical thoughts. This can be deduced from his letters and other biographical material. For instance, on 1 August 1781, Mozart wrote to his father Leopold concerning his living arrangements in Vienna, where he had recently moved:
Konrad cites a similar letter written from Paris, indicating that Mozart didn't compose where he was staying, but visited another home to borrow the keyboard instrument there. Similar evidence is found in early biographies based on Constanze Mozart's memories.
About 150 of Mozart's surviving works are incomplete, roughly a quarter of the total count of surviving works. A number of completed works can be shown (e.g. by inspecting watermarks or inks) to be completions of fragments that had long been left incomplete. These include the piano concertos K. 449, K. 488, K. 503, and K. 595, as well as the Clarinet Concerto K. 622.
It is not known why so many works were left incomplete. In a number of cases, the historical record shows that what Mozart thought was an opportunity for performance or sale evaporated during the course of composition. Braunbehrens (1990) observes: "Most pieces ... were written on request or with a specific performance in mind, if not for the composer's own use. Mozart frequently emphasized that he would never consider writing something for which there was no such occasion. Indeed, hardly a single work of his was not written for a particular occasion, or at least for use in his own concerts.
Mozart evidently had a prodigious ability to "compose on the spot"; that is, to improvise at the keyboard. This ability was apparent even in his childhood, as the Benedictine priest Placidus Scharl recalled:
As a teenager visiting Italy Mozart gave a concert in Venice (5 March 1771). According to a witness "An experienced musician gave him a fugue theme, which he worked out for more than an hour with such science, dexterity, harmony, and proper attention to rhythm, that even the greatest connoisseurs were astounded.
Mozart continued to improvise in public as an adult. For instance, the highly successful concert of 1787 in Prague that premiered his "Prague Symphony" concluded with a half-hour improvisation by the composer. For other instances, see Mozart's Berlin journey and Dora Stock.
There is apparently little evidence to bear on the question of whether Mozart's improvisations were a source of ideas to him for permanent compositions.
On one occasion, Mozart evidently used his improvisational ability to bolster his limitations in sight-reading. The composer André Grétry recalled:
The meeting of Grétry and the young Mozart apparently took place in 1766.
Braunbehrens suggests that on at least one occasion, Mozart met a deadline by simply not writing down part of the music and improvising it instead. This was evidently true of the Piano Concerto in D, K. 537, premiered 24 February 1788. In this work, the second movement opens with a solo passage for the pianist. The autograph (composer-written) score of the music gives the notes as follows:
Braunbehrens and other scholars infer that Mozart could not conceivably have opened a movement with a completely unadorned melody line, and instead improvised a suitable accompaniment for the left hand. Similar passages occur throughout the concerto.
The work was published only in 1794, three years after Mozart's death, and the publisher Johann André found some other composer (whose identity is unknown) to fill in the missing passages; it is these interpolations that have standardly been performed since that time.
Mozart appears to have possessed an excellent memory for music, though probably not the quasi-miraculous ability that has passed into legend. In particular, the use of keyboards and sketches to compose, noted above, would not have been necessary for a composer who possessed superhuman memory. Various anecdotes attest to Mozart's memory abilities.
Two of the violin sonatas gave rise to anecdotes to the effect that Mozart played the piano part at the premiere from memory, with only the violinist playing from the music. This is true for the Violin Sonata in G, K. 379/373a, where Mozart wrote in a letter to Leopold (8 April 1781) that he wrote out the violin part in an hour the night before the performance "but in order to be able to finish it, I only wrote out the accompaniment for Brunetti and retained my own part in my head. A similar story has survived concerning the Violin Sonata in B flat, K. 454, performed before the Emperor in the Kärntnertortheater 29 April, 1784.
One may perhaps question whether in these instances Mozart retained the entire keyboard part note for note in his head; given the independent testimony (above) for his ability to fill in gaps through improvisation, it would seem that Mozart could have done this as well in performing the violin sonatas.
Another instance of Mozart's powerful memory concerns his memorization and transcription of Gregorio Allegri's "Miserere" in the Sistine Chapel as a 14-year-old; for discussion, see Miserere (Allegri). Here again, various factors suggest great skill on Mozart's part, but not a superhuman miracle: the work in question is somewhat repetitive, and that Mozart was able to return to hear another performance, correcting his earlier errors. Solomon suggests that Mozart may have seen another copy earlier.
Konrad describes the views that were prevalent during the 19th century period of Mozart scholarship. In particular, "the 'making of music' was ... mythologized as a creative act". The 19th century regarded Mozart's compositional process as a form "of impulsive and improvisatorial composition ... an almost vegetative act of creation. Konrad states that the 19th century also mythologized Mozart's abilities in the area of musical memory.
An important source for earlier conceptions concerning Mozart's method of composition was the work of the early 19th century publisher Friedrich Rochlitz, who propagated a number of anecdotes about Mozart which were long taken to be authentic, but with more recent research are now widely doubted. Among other things Rochlitz published a letter, purporting to be by Mozart but now generally considered fraudulent, concerning his method of composition. This letter was taken as evidence concerning two points considered dubious by modern scholars. One is the idea that Mozart composed in a kind of passive mental process, letting the ideas simply come to him:
As evidence that the Rochlitz forgery does not provide an accurate picture of how Mozart himself perceived the act of composing, Konrad lists the first (authentic) quotation from Mozart given above.
Rochlitz's forged letter also was used in earlier study to bolster the (apparently false) story that Mozart could compose relying entirely on his memory, without the use of keyboard or sketches: