Monday, February 15, 2010

Atzenbrugg transformations, part 2

The six Atzenbrugg dances may well have originated in improvisation during the vacationers' evening dancing at Atzenbrugg Castle in July 1821 -- or they may have been occasional pieces composed during one of Schubert's morning sessions and then played later on in the day. Their somewhat advanced expressive qualities even suggest that they may have been played as a set in performance, rather than for dancing.

Each of the dances exhibits direct (chord-to-chord) mediant shifts. Each on its own is not extraordinary -- even as early as 1820 or 1821 -- but, taken together, they seem to me a remarkable hint at mediant play in Schubert's improvisational-/compositional- thinking at the time.

In n1 (D145n1), a fanfare-processional first phrase is immediately answered by a shift to the relative minor (R). From first strain to second, also, a P transformation, a hint of a linkage between different modes of efficient voice-leading on the "Riemannian hand."

In n2, the R move is complicated a bit more by the octaves but is again associated with a significant design articulation. Ditto the LP move between strains, and P for the second strain's latter half, which transposes the first strain's second half from C to A (an RP move if it were done directly).

In n3 (which is D365n29), the design/transformation alignment continues between the strains (chord roots D-F#; move is LP).
The non-tonic opening is not related to this pattern, but is of course striking in itself. It does announce a round of after-beat parsimony, however, depicted below with the chords of the reprise (change F# in the first and fifth bars to F-natural and you have the sequence of the first strain).