In most cases I have an overall idea of the direction of the music and how I envision it evolving - that can be a structural idea, a strictly sonoristic one but rarely a lick or melody. The latter are shaped afterwards and are what is perceived by the listener - that surface layer I mention in the post above. In this case I was thinking of ways to implement all 12 tones and therefore achieve a sort of emancipation from tonality but at the same time retain it's key characteristics, not the T-S-D-T cadences of course, rather it's gravitational pull towards an assigned center (in this case, a temporary one - the changing pedal point which, with it's consequent notes stacked on top of one another, forms the most tonally ambiguous diminished 7th chord). I find that produces an interesting effect, chaos withing order, and is what prevents that particular section from making you feel seasick, in other words, the listener is not overwhelmed by a barrage of atonality, although it is definitely not tonal at all.
To answer the second question - no, the B section is not in any way derived from section A. That's the idea, it's a contrast in almost every aspect possible - that of texture, rhythm, metric, melody and harmonic color. It is even quite classical in terms of periodic organisation - two phrases (antecedent and consequent) of equal length as opposed to the 4:3:2:1:0.5 asymmetrical relation of phrasal subdivisions in section A.
The motivation is similar to the one I presented above. The path is different, therefore the result too. Here I was looking for an ostinato pattern to support a free-form jazz improv allusion in the right hand. Again, think of it as a distant cousin that has inherited just the right features to still be recognized as a family descendant. The improvisational manner is very apparent - starts slow, phrases are more spread out; gradually intensifies; climax; cools down and restates the opening motives; dies away.
That is a cause and effect game there. If you look closely, the regions, although six, are really only two with three different transpositions starting on F#, Bb and D (notice the other tonally ambiguous chord implied here - the augmented one). Now, if you hold down the notes forming the ostinato pattern you'll notice that in one case you are left with a near whole-tone scale order and in other, with two groups of three chromatic notes a minor third apart. The linear contour of what you have at your disposal is quite different, therefore the melodies which stem out of it have a different shape. I find it interesting, rather than perplexing. The unity is still within the direction of the whole movement which I tried to depict.
[snip attached below] Each "deployment" of the row is contained within a rectangle. The different colors are to represent the varying sections. In red, it is stated three times exactly as it is conceived, in it's raw form - no additives. That's the exposition. Each time it is transposed a fourth apart (here I was thinking of the quartal harmony of the B section in the first movement. As extremely vague and distant as it may seem at this point, it gave me a creative impulse and that's what's important here. I do not delude myself that anybody could ever notice that unless stated. It wasn't intended to be obvious, it's merely one of these invisible laws that govern the piece). In the next section, "blue", the row is stretched to fill the same length of four bars by only appearing two times, as a result some notes are repeated. There is also an additional layer of 2-note sonorities remotely tied to the row itself but I will not go into detail here for the sake of brevity. Next is section "green" in widely spaced parallel major thirds - two unequally distributed rows as opposed to the symmetrical "blue" section.. In section four, the yellow one, the "theme" is diluted even further. Now it occupies the entire four measure expanse. The change of articulation, added embellishments and dynamic contrast is quite obvious and adds interest. What follows from here on are the same sections in reverse order - the whole movement has a mirrored form. The "red" section, here the last to appear at m.97-100, is a bit varied to serve as a build-up - the row is very muddy in the lowest register of the piano and is in it's reprisal form (three reiterations a fourth apart); a whole-tone cluster gradually develops on top.
I am glad you picked that up (note, here again in reverse - B followed by A)! I've gotten that before too, the reminiscence being short-lived. I agree. Let's not forget it's supposed to be a sonatina after all, not an actual sonata. I see it as a resume. It is quite rich in events but packed within a smaller box.
It was a good brain exercise for sure! It helps me understand better what I did by thinking of ways to verbalize it. Thank you for your questions.