Lovely Weekend Guest, a first-rate pianist and knowledgeable lover of Bach's keyboard works (she often refers to them by BWV number, for instance) spent part of her last weekend evening here in a friendly but, um, animated discussion with me of Glenn Gould's 1955 and 1981 recordings of Bach's Goldberg Variations (or BWV 988 as LWG calls it just to piss me off). On the whole, I rather prefer the earlier reading, she the later. The readings, both of them, are, of course, marvels, and display in equal measure Gould's trademark, irreproducible, and nonpareil Bach performance technique the preternatural (horizontal) delineation of multiple contrapuntal lines while maintaining fully their interlaced (vertical) harmonic coherence; the uncanny rhythmic sense, perfectly precise but infinitely plastic; and the equally precise and plastic articulation which commentators and critics insist on referring to as détaché but which is rather a near-perfect pianistic analogue of the highly prized and near-impossible harpsichord legato but are otherwise world's apart in spirit.
Rambunctious, LWG calls the earlier reading. Staid, I call the later, but only half mean it, using the term largely for reasons of symmetry with her rambunctious. About the only interpretive point on which we agree is that both Gould readings of this without-equal crown jewel of the entire Baroque keyboard repertoire blow away all other readings, truly excellent though some are.
I've had the recordings of both readings in my library almost since the day of their releases (both of which recordings survived the catastrophe that consumed most of the rest of my libraries several years ago), and though I listen to them often, I never really made the effort to nail down exactly why I preferred the one reading over the other. Better late than never, I decide, and so use the weekend's discussion with LWG as a spur.
Leaving aside the Aria (the ground bass of which is the unifying theme upon which the variations are built) which opens and closes the set of variations (and which I here leave aside mostly because I suspect I'm missing something important concerning it in the later reading as I simply cannot conceive a reasonable musical, aesthetic, or emotional justification for the funereal, almost structure-destroying tempo taken for it by Gould in that reading), I always vaguely imagined it was the generally slower — at times significantly slower — tempi of the later reading (i.e., slower as compared with the earlier reading) that provoked my antipathy for, even annoyance with, that reading. But I now see that's not it. Something else. And that something else is, I think, most clearly exemplified in Gould's two readings of Variation 25, dubbed "The Black Pearl" by the great harpsichordist and Bach scholar Wanda Landowska; an astonishing piece of music that enters the set of variations like some alien presence and at its departure leaves one slack-jawed with amazement and wonder.
With its frequent, expectation-subverting harmonic modulations, actual and implied, Variation 25's chromaticism is so extreme it makes the music of Tristan and Parsifal seem positively diatonic by comparison. Gould, in his liner notes for the 1955 release, refers to this variation in his typically provocative way as a "Chopinesque mood-piece," and elsewhere as a "Romantic effusion." It's neither of these, of course, but thoroughly and essentially Bachian throughout. It's simply that it seems to not belong to this set of variations but seemingly magically still remains one with it while at the same time seeming to inhabit another world altogether.
In his earlier reading, Gould captures this another-world quality perfectly, largely by his circumspect and strategic use of rubato, an expressive device more identified with the Classical period and with Chopin and the Romantics than with Bach but here used to brilliant effect. The use of rubati in this earlier reading makes the progress of the melodic line and the movement of the astonishing harmonic shifts seem to play out on a scale cosmically slow the variation's internal pulse seemingly proceeding as do the slow wheeling of galaxies through the infinite eons of boundless space, the harmony's expectation-subverting modulations given point by the rubati and suggesting tonalities mystically strange and vastly remote.
As I said, music that leaves one slack-jawed with amazement and wonder — in the earlier reading, that is. In the later reading the rubati are more temperate, the internal pulse and inflection of the melodic line and harmonic shifts more deliberate and sober, the variation now seeming more at home with the set's other variations than in the earlier reading because more metrically and sensibly proportional to them. The expectation-subverting modulations are still there, of course, and still give this variation the feel of a vaguely alien presence within the set, but the another-world magic of the earlier reading has gone missing. Gould's deliberate and sober approach to the music bars such mysteries. One's amazement at the variation's harmonic inventiveness remains, of course, but the wonder has fled.
This stately, deliberate sobriety characterizes most (but not all) of Gould's later reading of the Goldberg; a sobriety achieved by various musical means, one of which is Gould's singling out of one of the contrapuntal voices (typically, the unifying element of the Aria's bass line) as if to say, "See? Here it is." In intent, the device is somewhat akin to that used in "Konzept" opera productions wherein the self-involved director insists on pushing his particular "vision" repeatedly in the audience's face lest they miss it. It's a device as annoying here as it is there.
The sheer audacity the "rambunctious[ness]" of Gould's earlier reading may be considered by some a mark against it, but for me it's that very audacity that gives revelatory life to that earlier reading, while the deliberately sober later reading largely (but, again, not entirely) eschews audacity as if to act as corrective of the implied youthful excesses of the earlier. While the later reading may achieve its end in that respect, it does so at a cost, and to my way of thinking (and hearing), the cost is simply too great.