The idea being, that Schoenberg first composed his Op.47 Phantasy as a violin unaccompanied piece, and subsequently added a piano accompaniment. Now, perhaps this was strictly true; and if I erred in my own long-gestated notion, it was in supposing that there was a greater distance of time and artistic intent, between the violin unaccompanied phase, and the let's add piano accompaniment phase.
Even if that thought is, regarding the Schoenberg piece, technically an error, I have made artistic use of the idea, and put it into practice with the new quartet version of the Angular Whimsies.
Today I found an interesting article on the Schoenberg Op.47. One incidental bit of interest is, that it was published the same spring that I was graduated from the College of Wooster . . . and it thus predates my Buffalo studies. In the scope of this paper, the matter is not settled; the author allows the possibility that at no point was it actually a violin solo piece; and delicately observes that Josef Rufer does not explain how it is known that the violin part was composed first.
In reading the article, I found what strikes me as an apparent misprision (but it may be simply that I have been somewhere inattentive, myself) on the part of the author. In his Example 7, he gives Schoenberg's row divided into two hexachords:
Bb - A - C# - B - F - G / D# - E - C - D - Ab - Gb
10 - 9 - 1 - 11 - 5 -7 / 3 - 4 - 0 - 2 - 8 - 6
He correctly points out that the second hexachord is a transposed inversion of the first.
He then reorders the pitches of the first hexachord, so that the discussion is not solely of contiguous chains of pitch-classes within the strict series . . . and this is where (I think) he offers a remark which I think an error:
It is significant that the hexachord contains no major or minor (...) triads (....)
Perhaps his eye was misled by the spelling of C# in the first hexachord, and of D# in the second; but respelling them Db and Eb respectively, then the bb minor triad is a subset of the first hexachord, and a corresponding major triad is necessarily a subset of the inverted second hexachord, in this case, Ab major.
It may be that Schoenberg does not employ them as such, but in fact the hexachords do contain a minor and a major triad.