I had written this note and then discovered that a month ago Denis Fodor
posted excerpts from a Times Literary Supplement review of the book
mentioned below:

In the last couple of days, in conjunction with beginning 'The Compleat
Brahms', edited by Leon Botstein, I've been listening to a lot of Brahms
piano music.  I'd always much preferred the orchestral music to that for
the piano alone.  Partly, I think, I felt that way because in my playing
days I'd not especially enjoyed playing Brahms myself on the piano.  I'd
though it was too bass-dominated, too muddy.  And, I blush to say, not
showy enough.

But I reached for an old Stephen Kovacevich recording of most of the late
pieces (Opp.  116, 117, 119) and found myself transfixed.  Partly by the
playing, but mostly by the music.

So I got out the scores and began fumbling around at the piano.  My
technique is shot to hell and my arthritic hands hurt, but I'm sort of
making music, and perhaps more important, making discoveries.  What tidy
little pieces these are!  Interconnected thematically and harmonically in
ways I'd never realized; I can thank Botstein's essayists for some of that
insight, but also under my fingers the discoveries abound.  One hears,
for instance, that firm tonality goes awryin some of these pieces.  And
there is a method in his madness, for instance, in the Op.  116 Fantasies;
tonality starts fairly conventionally in the first piece, goes more and
more adrift as we go towards No.  4, and then becomes more conventional as
we progress out to the last piece, No.  7--a kind of arch form of tonality.
No wonder Schoenberg loved this music.

I must say it's a delight to be able to discover something new in music
I'd thought I knew well.  It turns out I only thought I knew it.  In his
preface Botstein makes the point that the 19th-century music lover knew
music at a much more basic level because she primarily had make it herself
by playing it on the piano or other instrument.

He also emphasizes that there is too much current emphasis on
establishing a canon of *recordings* when it's really a canon of *works*
that is important; focus on recordings tends to move current performers'
interpretations away from the score and toward prior recorded performances.
Hence, he makes no references to recordings of Brahms' music.

Once again I've demonstrated to myself the limitless delights lurking below
the surface of thrice-familiar music.

Scott