This is a pretty eclectic bunch of recordings. Probably my favorites in this cluster are #35 -“Consort of Musicke Bye William Byrde and Orlando Gibbons,” followed by #38 – the Handel Suites, as a close second. The three Mozart recordings are tied for worst in this group.
27. Mozart / Piano Sonatas, Vol. 1: Nos. 1-5, K. 279-283 (1967/1968)
These are the very earliest piano sonatas by Mozart – I read an interview in which Gould said these were his favorite. Even so, I just don’t get very much out of these recordings. I actually spent about an hour reading deep into an internet message board flame war about this (though now I can’t find it… oh well – it was very amusing reading if you like that sort of thing). Gould is widely reputed to have disliked Mozart, and said things like “he died too late, rather than too early” or something to that effect. The big argument is between the pro-Gould people who say his performances just reveal that banality of Mozart’s piano compositions, and the anti-Gould people who say he just had a lack of imagination and empathy with this music and didn’t know how to perform it well. Either way, the Mozart recordings aren’t my favorite (though the Fantasia and Fugue [K. 394] that’s in the much earlier Mozart/Haydn record is a pretty engrossing exception).
28. J. S. Bach / The Well-Tempered Clavier, Book Two, in three volumes (1966–67, 1969, 1971/1968, 1970, 1971) 
I haven’t invested the necessary time in all of this music. It’s 48 pieces: one prelude and one fugue in each of the 24 major and minor keys, just like the Well-Tempered Clavier Book One. My highly subjective and ill-informed impression is that the second Book (or at least, Gould’s second book) was more enjoyable than the first. But when I try to listen to either from beginning to end, have this weird OCD reaction that keeps me from reacting to it holistically. They weren’t recorded to be heard as a whole I don’t think – #24 in B minor is no more final than #1, in C major, as far as I can tell. They’re not really a coherent performance program like a four-movement sonata or three-movement concerto might be.
This was interesting just for the composers selected. Gould didn’t record very much of this sort of late-romantic/more tonal 20th century music. Both recordings have a distinctly sort of impressionistic feel to them, and are deeply redolent (for me anyway) of the feeling you get from walking through busy city streets, or would have 60-70 years ago. It doesn’t sound like Jazz but there are some Jazz-feeling harmonies here. Gould certainly succeeds in evoking a particular feeling (no idea if that’s what he was trying to do).
30. Mozart / Piano Sonatas, Vol. 2: No. 6, 7, and 9, K. 284, 309, 311 (1968/1969) 
See above – nothing really to add.
31. J. S. Bach / Clavier Concertos, Vol. 2: Nos. 2 and 4; Vladimir Golschmann / Columbia Symphony (1969/1969) 
This is the counterpart to two earlier recordings: Bach Clavier Concerto No. 1 (on the back side of Gould’s recording of Beethoven’s 2nd piano concerto), and also a later record that has no. 3, 5 and 7 on it. I prefer 1/3/5/7 to 2/4, for whatever reason (probably just familiarity). I still think this is one of Gould’s strongest areas though – Bach-era concerto. There is a magic to the interplay between the solo instrument and the relatively small backing orchestra (I think it’s only a few pieces, certainly not what you might picture when you hear “orchestra”) that allows the piano to take center stage the entire time, and the other instruments simply to color its movements. There is an undeniably energy on these recordings that at times almost turns the piano into a human voice, even as it’s involved in furious counterpoint.
The piano here felt primarily like accompaniment, so I found myself thinking more about the underlying music and less about the piano performance. It is of a piece with Brahms quartets I’ve listened to over the years, very warm and emotionally engaging.
33. Beethoven / Piano Sonatas, Op. 13 “Pathétique”, 27 No. 2 “Moonlight”, and 57 “Appassionata” (1960, 1966-67/1970) 
I’ve listened to these three sonatas on a Vladimir Ashkenazy recording so many times that it’s a little tough to hear Gould’s take as anything but wrong. Still, however, I do appreciate his “Moonlight”, since it’s actually louder and faster than it’s normally played, somehow allowing the chord textures to speak more clearly, since it takes away the sometimes ponderous emotional subtext this piece can summon up if played too melodramatically. In an interview Gould confessed that, though as a professional musician, he knew he should have more sophisticated opinions, he actually like Beethoven’s Moonlight Sonata most of all of Beethoven’s solo piano compositions. As far as Gould’s recordings of them go, though, I’m still most entranced by his second record – Beethoven’s sonatas 30-32, especially the second movement of 30, and the entirety of 32, which has to tendency to evoke such a sense of finality that you almost wonder how people could have persisted in sonata writing after hearing that piece – you almost want to say – that’s the final word on that.
Something about Beethoven’s overall approach seems to me to make the “variations on a theme” here not really work for him. Like it’s just in a style that’s suited to this sort of repetition (that said, the second movement of Beethoven’s 5th symphony is, to me anyway, one of the most moving examples of the form). I can hear Gould’s ambivalence about Beethoven in this recording.
This was a real inspiration. I’m really unfamiliar with Renaissance/Early Baroque, but this set of shorter pieces (probably not originally written for the piano) were hauntingly evocative. Ever since I was a little kid, going to church around Christmas, I’ve been deeply moved by “What Child is This?” or other hymns set to the same Greensleeves tune – it’s often so chilling I can’t sing. The pre-classical harmonies in this hymn are strangely disconcerting and at the same time reassuringly familiar. All of the pieces here recorded have that same quality – it’s so much more perplexing than the simple 1-3-1 cadences that you come to expect from the end of, say, a Mozart symphony. Try it yourself.
36. Mozart / Piano Sonatas, Vol. 3: Nos. 8, 10, 12, and 13, K. 310, 330, 332, and 333 (1955–56, 1969-70/1972) 
See above – except – the recording of No. 10 is interesting only insofar as it’s much faster and less deliberate than Gould’s recording of the same piece on his earlier Mozart/Haydn record.
37. Schoenberg / Songs, Op. 3, 6, 12, 14, and 48, and Op. posth., with Donald Gramm, Helen Vanni, and Cornelis Opthof, baritone (1964–65, 1968, 1970-71/1972) Handel / Suites 1-4 on the harpsichord (1972/1972) 
I definitely have not digested these songs. The earlier ones have that late-romantic German feel, and the later ones move towards the atonal. I wonder about atonal singing – it seems like it would be harder than atonal clarinet playing, for example, since you have to train yourself to sing non-melodic notes, and I imagine anyway that a lot of what a singer does to keep in tune has to do with tonal resonances that you just can’t take advantage of with this sort of music.
Probably my second favorite in this batch. This is the only thing Gould recorded on a harpsichord, as far as I know. Something he often complained about in interviews was the 19th century classical and romantic notion of dynamics – that everything’s always supposed to have these dramatic louder and softer spots, to charge the emotions at certain times. You can’t do that on a harpsichord (or an organ – like Gould uses on his “Art of the Fugue” earlier recording) which means you instead get what is apparently referred to as “terraced dynamics” – a whole movement is loud, then the whole next movement is soft (usually by having more or fewer instruments involved). Both a harpsichord and an organ are digital instruments – each key is either struck or not struck (just two states). I wonder what Gould could have done with a synthesizer – it’s surprising he never did anything with one. A piano is much more analog, in the sense that you can get an infinite variety of volumes out of it. There are moments on this Handel recording where it sounds like you actually are listening to a synthesizer – and I mean this in a good way. It’s wonderful that it IS human and feels quite human, but at the same time is regimented in a way that reminds you of a computer. I found that quality especially apparent in the 4th movement of the 1st suite (a really invigorating folk dance form), and in the 4th movement of the 3rd (which is a theme-and-variations sub-suite coming before the 5th movement. And since it’s a harpsichord of course, all of it reminds me of Rushmore, which is probably the real reason I like this record.