If the readings from ^8 (presented in yesterday's post) are problematic, still it is certainly true that Schubert puts plainly before our ears the tonal space of the upper fourth, ^5 to ^8, and we might reasonably engage that as a priority in the large-scale design. The first reading built on this basis is a simple rising Urlinie ^5-^6-^7-^8 that trumps Schachter's line from ^3 in simplicity and elegance: not only is the line short but the reading also rests on what we hear readily: "the right hand plays two melodic lines written in free imitation"; furthermore, the upper voice is the lead melody, even if that voice is "generally [less] active" than the alto (Schachter, 70).
We can also appeal to style statistics: although complex compound melodies exist in abundance in the violinistic Ländler repertoire, and certainly find their way into Schubert's waltzes (D145n1 (Ländler), D146n2, D779n26, D779n29, D924n3, and D969n10 are particularly nice examples), a middle-voice main melody is rare in the early waltz repertoire (although we could, I suppose, just declare this to be yet another of the anomalies in the A Major Waltz). In D734n14, for example, "soprano" and "alto" start an octave apart; the role of principal voice is initially contested, but the soprano wins out by bar 3. D779n1 (score reproduced at the bottom of this post) and D779n19 are the only other Schubert waltzes I know in which a really convincing argument can be made for alto-priority.
The alto replicates Schachter's main melodic voice throughout. The soprano also traces a middleground form of the line from ^5 to ^8 in the first strain (not shown). The ^5 is reinstated in the upper voice in the second strain, allowing the final upward-pushing figure ^5-^6-^7 over V, in a replication of the cadence of the first strain.
Although the specific figure Schubert uses here (both ^6 and ^7 over V) seems to have been an innovation of his, play with scale degree ^6 above both I and V was a cliché of the Ländler almost from the beginning and passed into the waltz once its subgenres began to merge in the 1820s.
The other reading focused on the space ^5 to ^8 is decidedly less successful. The symmetrical treatment of the space from ^5 to ^8 takes the strongest elements of the first 8-line reading and replaces its weak ending with the rising line of the first reading above. This version puts before the eye the aurally salient parallel patterns of rising in the first and second strains-the first time through the long initial arpeggiation, the second time through the ascent of the Urlinie to ^8 -- but it suppresses the same figures in the opening of the second strain (once again, however, that could be ascribed to differences of function -- or, here, function in different structural levels). I find the parallelism appealing, but finally prefer the previous reading (^5 rising to ^8) because it puts such emphasis on an unusual stylistic feature of this waltz: its play of counterpoint.
Reference:
Schachter, Carl. "Rhythm and Linear Analysis: Durational Reduction." Music Forum 5 (1980): 197-232. Reprinted as "Durational Reduction" in Schachter 1999a, 54-78. (Schachter, Carl. Joseph Straus, ed. Unfoldings: Essays in Schenkerian Theory and Analysis. New York/London: Oxford University Press, 1999).
(Score of D779n1 (click the thumbnail for the original image))
(Score of D779n1 (click the thumbnail for the original image))