This is a very interesting piece as it only really flirts with and teases one's sense of tonality. In addition, it moves quickly, catching the listener off guard.
The first four measures are a sort of vamping introduction with a g and an f sharp together on the first beat (the piece is in 2/4) and a g an octave down on the second beat. The f then becomes natural as the melody in the right hand begins, giving the listener a sudden jolt--for it really does seem almost tonal for a moment. The right hand makes one eight measure phrase and ends on a half cadence on G the dominant(we actually turn out to be in c minor, although the key signature marks no accidentals). After a pickup from the last bar of the first phrase and then six measures, we come to an almost deceptive feel which then resolves in the right hand to more a tonic -- running up the scale with c on the first beat. However, in the left hand, the tonality does not fit at all with c minor. By measure twenty five, we have a repeating G in the right hand voice, which could function as insinuating a dominant, or even a progressive modulation into G, but underneath in the left hand, we have an e, a g# and a c# that really wash out any c minor or G major that might have been inferred. These last nine bars of A plus pickup represent for me a transitional section (as that c# turns into a d flat) that perhaps elides with a sense of closure for the A section.
In B we are now in Ab major and the temp picks up a little. Also, he switches off the voice in a little contrapuntal immitation. The two voices (each hand only playing one line) begin to interact more closely until they unite together in a unison run all the way to the bar of 3/4 and the return of G in the bar following. I could almost divide this up into two phrase groups that any real cadence, but suggest an IAC after 8 measures, and then a HC after the next ten measures--thus making this section open.
Again we have four measures of introduction, though up an octave with the right hand playing out a syncopation on the g. The left hand plays much higher this time, and there are now sixteenth note runs in the right hand. In the thirteenth measure of this A' section, both hands join in complimentary 16th note runs, and suddenly change texture completely in bar 16. There the dynamic is marked forte, the tempo is up (as it was even slower than the first A originally), and we have this terminative section that establishes the feel of c minor firmly though in a rather unorthodox sense, ending on an open fifth on G. So A' is also open, ending on a half cadence.
The section divisions are marked with an added bar of 3/4 (both at the end of A and B) and through tempo markings. A seems to not only be open in the sense that the right hand sung out a G until the end of the bar, but that c# functioned as a d flat for the B section, giving it a HC feel in the Ab major as well. Also the differences in texture make each section distinct. As for A' ending in the original key, at least it does not play the c# so at least leaves us with a HC feel in the original key. In A', I cannot really seem to find cadences or phrase groups, though the score is divisible through texture changes and tempo markings. Because the sections are small and lack higher organization in themselves, I see this as a simple ternary in which all sections are open (though B and A' end on their own dominant, and A seems to want to mix both the dominants of B and A').