This movement has gone, over the last weeks, from being a piece that I’ve long admired to something of an analytic enigma. Simply put, I am at odds to determine whether this “Rondo-Finale” is best considered as a rondo, a sonata-allegro or a fugue.
First, to the title, if that can be a clue for the analyst. Does “Rondo-Finale” suggest “rondo-as-finale” or “rondo-then-finale?” My hearing suggests that there is indeed a rondo here, and that it is followed by a lengthy coda, so that the second possibility seems stronger. In this case, the coda could perhaps begin in m. 581, at a key change to A-flat major, far-removed from the home key of D major. The melodic and motivic material is related to the rondo theme (mm. 24-55), but this late harmonic move away from the home key suggests a coda.
The music up to this point is highly suggestive of rondo technique, specifically of five-part rondo with its three statements of the rondo theme with interspersed refrains. The second refrain (beginning in m. 167, the “C” of “ABACA”) is the longest, and is heavily reliant on developmental techniques, especially exploration of remote key areas and contrapuntal recombination of motivic material.
It is, however, the first refrain (beginning at m. 56, the “B” section) that is most striking. It suggests a four-part fugal exposition, first with a running eighth-note subject, then with various countersubjects introduced over the eighty bars of this section before the return of the rondo theme. The second refrain can then be cast as a continuation of the fugue. At m. 273, the original fugue subject appears in counterpoint with one of the countersubjects. This countersubject becomes the second subject of a double fugue that dominates much of the rest of the second refrain (development). Contrapuntal technique abounds, with the inversion of the second subject appearing in the violins at m. 457.
The final entrance of the rondo theme appears in a highly modified form at m. 497, leading not back to the beginning, then, but toward the “finale” section of the movement.
There are also intimations here of sonata-allegro, or at least something along the lines of a hybrid sonata-rondo, as found in another wonderfully contrapuntal work, the finale of Robert Schumann’s Piano Quintet, Op. 44. The first rondo and first refrain would correspond to a sonata exposition, the second rondo and second refrain to the development, and the last refrain and “finale” to the recapitulation and coda.
It is a fascinating feature of this movement that it not only is a highly compelling piece of music, but that it also embodies these three formal procedures .
A final issue with this symphony is its harmonic plan. The home keys of the five movements are, in order, C-sharp minor, A minor, D major, F major and D major. Mahler uses this narrative tonality in other places. If Mahler had in mind a major-key finale to a minor-key symphony, in the manner of Beethoven’s Fifth and Ninth Symphonies, perhaps he felt that the technical challenges involved in C-sharp major might be too much for his orchestra, given the already stiff demands of the music. At any rate, it is also simply possible that Mahler is moving away from the single-key concept of a symphony. The five movements appear appear to be held together from a motivic standpoint, rather than from harmonic consistency, but in a traditional sense, they are no more related than a suite of pieces extracted from an opera or ballet. It is a testament to Mahler’s compositional technique that the piece feels completely unified without sharing a common key center.
Now on to the Sixth–four large movements, so two weeks each.