Schubert: Piano Sonatas D.959 & D.279
Wilhelm Kempff, piano
Deutsche Grammophon 2530 327
Everything Franz Schubert ever did often sounds like different sets of themes and variations–for quartets, orchestras, and soloists alike. But listening to Franz Schubert’s late piano works bears out that theory. Each one is like overhearing the composer plan, rework and deconstruct his music. The sonatas D.958-D.960 (nos. 19-21) are hallmarks of the form, thematically tight and near minimalist in their reiterative structure. In the D.959 sonata and elsewhere, a simple theme is laid out, evolved subtly or in startling shifts and transpositions; troubling tonal conflicts intrude, with flats and sharps thrown in like minefields scattered in the score. As with the other sonatas in the cycle, it’s a metamorphic experience.
Finished in 1828, the D.959 sonata has the distinction of containing one of the most intensely unsettled movements in the canon. You know something is off with the Andantino second movement by its key signature, F major, when the other four movements are in A major. It’s a creeping surreal nightmare in between moments of the uneasiest calm.
Any preference for a particular performer, for me, is based on the Andantino, as it encapsulates so much of what Schubert was all about. Valery Afanassiev’s ECM rendition is my go-to: meditative, agonizingly measured, almost all-holds-barred, and sounds a little like Andrei Tarkovsky switched from filmmaking to pianism. Rudolf Serkin’s vinyl recording from Columbia is an edge-of-your-seat performances, while Krystian Zimerman’s recent release (destined to be compared to the greats, and kindly pressed for audiophiles) is totally uncluttered and the probably the least ostentatious version I’ve ever heard yet.
Here, Wilhelm Kempff takes a middle-of-the-road approach that seems spot-on and sticks close to the original. Dramatic where it should be; delicate in the lyrical moments. My one complaint is his cutting some of the repeats in the Allegro, which comes in at about 4-5 minutes shorter than other recordings. The D.959 is accompanied with the early D.279, a precocious, unfinished-seeming piece that merely hints at later volatilities.