Saturday, July 11, 2015

Retro Record Review #5: Bach, Goldberg Variations

This is an occasional, very occasional, recurring item where I review albums that might have been issued quite a while ago, but are still important. I haven't done one for nearly a year, so overdue. This one is prompted by the Guardian's Face the Music series where they get musicians to answer a series of questions--always the same questions--and call it an interview. Cheap to do, anyway. One of the questions is always "how many versions of the Goldberg Variations do you own?" The answers are always, well, less than interesting. The last answer, from a fairly young composer, was the Glenn Gould first recording from 1955. Yes, a stupendous debut. The Goldbergs were also the last recording he made before his death in 1982, so the two recordings are like bookends to his entire career. Along with Gould, I rather prefer the second recording. But citing either one is not too imaginative, now is it?
How many recordings of the Goldberg Variations do you own? Do you have a favourite?
Only one: the 1955 Glenn Gould recording of the Goldberg Variations. It is hard to listen to any other interpretation.
Hard to listen to any other interpretation if you have only been listening to Glenn Gould, I would imagine. So today I am going to review three different recordings (none by Glenn Gould) of the Goldberg Variations, published in 1741, during Bach's lifetime, unlike many of his compositions. The title page says quite clearly:
Clavier Ubung / bestehend / in einer ARIA / mit verschiedenen Verænderungen / vors Clavicimbal / mit 2 Manualen. / Denen Liebhabern zur Gemüths- / Ergetzung verfertiget von / Johann Sebastian Bach / Königl. Pohl. u. Churfl. Sæchs. Hoff- / Compositeur, Capellmeister, u. Directore / Chori Musici in Leipzig. / Nürnberg in Verlegung / Balthasar Schmids.
Which translates as:
Keyboard exercise, consisting of an ARIA with diverse variations for harpsichord with two manuals. Composed for connoisseurs, for the refreshment of their spirits, by Johann Sebastian Bach, composer for the royal court of Poland and the Electoral court of Saxony, Kapellmeister and Director of Choral Music in Leipzig. Nuremberg, Balthasar Schmid, publisher.
We are going to need the score, which you can download here. You can even get a facsimile of the first edition. Here is the Aria that the 30 variations are based on. We learn from Christoph Wolff's excellent book on Bach that the Aria is based on a ground bass used by Handel in his Chaconne avec 62 variations, published in 1733. Bach found this simple ground bass to have rich possibilities which the Goldberg Variations certainly demonstrate.

As this was written for the harpsichord, not the piano, I think I will just review some recordings of it on harpsichord, those by Gustav Leonhardt, Trevor Pinnock and Scott Ross. The oldest recording is that of Gustav Leonhardt recorded by Harmonio Mundi in 1978 (my copy came in this box set):

Every third variation is a canon and each one is at a steadily increasing interval. The first is at the unison, the second at the second and the third at the interval of a third. But as you can see, the second voice enters a third below the first, not above:

The canon at the 4th is not only a 4th below, but in inversion as well, meaning that all the intervals are inverted; where the dux (first voice) goes up, the comes (second voice) goes down and vice versa:

He reverses this in the second half with the comes above the dux. The canons continue expanding until the 27th variation at the interval of a 9th. Variation 26 has the very unusual time signature of 18/16 in the top voice with 3/4 in the bottom, which is the equivalent of notating the upper voice in sextuplets. And, of course, this reverses from time to time with the 3/4 above and the 18/16 below.

Leonhardt's performance is the shortest of the three, though not because of fast tempos. A lot of his tempos are a bit more relaxed than the other two. The reason for the shorter recording time is that he takes none of the repeats. If you do that, you can play the whole set in under 50 minutes: 47:19 in Leonhardt's recording. As this was made before the advent of the CD, I suppose it might be because it was very difficult to fit them all on a single LP with the repeats.

Moving on to Trevor Pinnock, his performance was recorded in Paris in 1980 and my copy came contained in this box set. The CD contains the Italian Concerto as well as the Goldbergs. The total duration of the Goldbergs, minus the Italian Concerto, is 60:53 in Pinnock's recording according to my calculation. Now, unless Pinnock is taking some really superhuman tempi, this is still not long enough. The reason is that, while Leonhardt takes none of the repeats, Pinnock just takes some of them. Neither half of the aria is repeated, but both halves of the first and second variations are, for example--but not the third variation, so what he repeats seems a bit random. They don't tell you what instruments are used, but the sound of Pinnock's harpsichord is a bit warmer than Leonhardt's and the recording ambiance is a bit more resonant. Some of Pinnock's tempi, such as for the gigue-like 7th variation, are brisker than Leonhardt's, but others, such as Variation 9, the canon at the 3rd, are slower. Incidentally, if you are looking at a recording of the variations in a store and wondering if the repeats are taken, if they are, the duration for most of the variations will be between 1:30 and 2 minute. If they are not taken, you will see timings under a minute. While Leonhardt's recording is not lacking in virtuosity, in some of the variations, such as #14 with its scampering 32nd notes Pinnock's is more striking in its brilliance:

This is even more evident in the very challenging 20th Variation with its high-velocity 32nd note triplets that Pinnock delivers with great aplomb:

The expressive heart of the Goldbergs is Variation 25, an adagio that journeys to some very strange places:

The last of the three recordings is that of Scott Ross, recorded in Ottawa in 1985.

This was reissued by Erato in 1997 and is still available. Ross' Aria is the most spritely of the three and many of his tempi are just a bit faster than the previous two players. His total duration for the variations is 69:31 which means that he is taking all the repeats. He has perhaps the greatest technical ease with the music and the interpretations verge on the insouciant at times, something you rarely hear in Bach. Occasionally you might prefer a slightly slower tempo as in Variation 13. But when it comes to ones like Variation 14, they really benefit from the energy. Perhaps a better word to describe Ross' Bach rather than insouciant is sprezzatura--as if you are playing with no effort, with complete ease. And so it is. This recording was made just after he had completed his recording, over eighteen months, of all 555 sonatas by Scarlatti. The recording includes a filler of the Scarlatti Sonata K 146. But wait, it's not a filler, it's an encore for we discover at the end of the Goldbergs that this is not a studio recording, but a concert recording, with audience. I don't know if they edited the performance afterward, something I know you can do and, if you do it immediately after the concert in the same hall, your edits will be unobtrusive as they will have the same acoustic ambience. But, this could be unedited, just like the recordings of Grigory Sokolov. In either case, a remarkable performance.

So, who do we prefer? All three of these recordings were made in a fairly brief period between 1978 and 1985 when the harpsichord, as part of the Early Music Movement, was really coming into its own. At this point harpsichordists were finally ready to take back the most virtuoso pieces in their repertoire, ones that had been "owned" for a long time by pianists. Indeed, in the notes to the Ross recording he makes some rather unkind remarks about Gould's Bach.

All three of these recordings are very fine, each preferable in their own way as each musician is a great artist. The Leonhardt is perhaps the most expansive, which sometimes is to be preferred. The Pinnock has moments of expressive depth and moments of virtuoso brilliance. And Scott Ross' performance has all the verve, rhythmic clarity and elegance that we would expect. None of these recordings disappoint in any way.

But perhaps none of them achieve quite the meditative interiority that Gould does in his second recording, especially in that opening Aria:


A.C. Douglas said...

"It is hard to listen to any other interpretation [after Gould's]."

It is, you know. Somehow they all sound "wrong". Very wrong, Landowska's included.


Bryan Townsend said...

Well, there are reasons why Gould's two recordings are so very famous and popular. Good reasons. But my feeling as a performer is that no-one ever "owns" a piece like the Goldbergs, not even Gould. I hear a lot of good things in the three harpsichord versions. I think that people listen so much to Gould's Goldbergs that they get imprinted. Doesn't mean the other performances are wrong, just that you are over-exposed to Gould.

A.C. Douglas said...
This comment has been removed by the author.
A.C. Douglas said...

"I think that people listen so much to Gould's Goldbergs that they get imprinted."

That may well be true generally and I'm of course well aware of that. But it runs deeper than that, I think. It does so in my case at least. Whenever I listen to Gould's reading I simply can't shake the impression Gould had a direct line, so to speak, which line runs: Gould to Bach's deepest internal vision of the work, and Bach to God's.

Sorry. Didn't mean to go all metaphysical on you. Seems an ineluctable hazard whenever one speaks of encountering this work and Gould's realization of it.


Bryan Townsend said...

I know exactly what you mean! There are so many things about music that we cannot express in words. I haven't been listening to Gould's Goldbergs recently, but I think I feel the same about some of his Well-Tempered Clavier. There is a clip on YouTube of Gould playing the E major fugue from Bk 2 in an absolutely transcendental way that no-one else has even come close to.

A.C. Douglas said...

Some years ago I had a flash of insight concerning Gould's Bach readings generally and wrote the following in a 2006 S&F entry:

=== Begin Quote ===
Listening last night to Glenn Gould's reading of The Well-tempered Clavier, Books I and II, I wondered, and not for the first time, what it was about Gould's Bach that made it so compelling, even — dare I say it? — transcendent. There are a number of readings of Bach's keyboard works by harpsichordists and pianists that are by any fair-minded and honest assessment first-rate ... but no other reading of my experience possesses that uncanny quality of almost preternatural rightness that's the preeminent hallmark of Gould's Bach readings.

Perhaps the most illuminating nontechnical characterization of Bach's keyboard polyphony — illuminating for performer and listener alike — is that, at bottom, it's an in-progress intellectual and philosophic conversation carried on by intellectual equals. And, interestingly, embedded within that anthropomorphic characterization lies, I think, the answer to the question.

While all first-rate keyboardists recognize that in-progress conversation and acknowledge its existence in their Bach readings, Gould alone among them understands precisely what each speaker is saying, knows exactly what the conversation is about, and understands fully all its manifold implications.
=== End Quote ===

Thought I'd share that.


Bryan Townsend said...

Very, very interesting thought and very well put!

Bryan Townsend said...

Have you heard Grigory Sokolov's Art of Fugue?

A.C. Douglas said...

No. I don't know Grigory Sokolov's work at all. Not that surprising, actually. While Sokolov's name is of course known to me, I've a marked antipathy for the piano as an instrument, generally speaking, and have very little in my existing library featuring that instrument.