The story behind the creation of the work, as told by an early Bach biographer, is utterly charming. In Dresden, one Count Keyserlingk employed a young house musician, Johann Gottlieb Goldberg. The Count

“was often unwell and then had sleepless nights. On these occasions, Goldberg had to spend the night in an adjoining room so that he could play something to him during this sleeplessness. The Count once remarked in Bach’s presence that he would very much like to have some keyboard pieces for his Goldberg, of a character so gentle and somewhat merry that he could be a little cheered up by them in his sleepless nights. Bach believed that he could best fulfil this wish with some variations, which until then he had held to be a thankless task because of the basic harmony always being the same.”

(Unfortunately, the story is probably inaccurate. The piece contains no dedication to the Count, and Goldberg was only 13 years of age in 1739-40, when the music was probably written. More likely, Bach actually wrote the piece for his son Wilhelm Friedemann (hence the phrase "Keyboard Practice") and presented a copy to the Count when Bach visited Wilhelm Friedemann (who, misunderstanding or misremembering many years later, was probably the source of the story)in Dresden in 1741. One may speculate that Bach did so in the hope of winning future commissions from the Count in order to supplement his income. If so, Bach was disappointed.)

Technically, the music is an extraordinary tour de force. Consisting of 30 variations on an original theme, the music explores a wide variety of contrapuntal devices, most rigorously the canon (a form of strict imitation resembling, at its most basic, a round like "Row, Row, Row Your Boat"). Canons put in regularly scheduled appearances at every third variation – one musicologist likened its structure to a string of rosary beads – with each canon beginning on a successively higher pitch. To add to the complication, some of these canons are inverted (the imitation is upside down rather than literal).

I generally prefer to hear Bach keyboard works played on the harpsichord, but the Goldberg Variations is the exception. The clear first choice, in my book, is Glenn Gould’s 1955 version on piano. This recording put Gould, then 22 years of age and virtually unknown, on the map; it also put the Goldberg Variations on the map. I’ve run across a description of the impact of Gould’s recording that’s so perfect that I’m just going to quote it:

“Until Gould, the piece, when played in public at all, was largely performed by Baroque specialists, usually on the harpsichord, and often presented in a ‘now you must take your medicine, it's good for you’ spirit. Listening seemed a character-building chore rather than a pleasure. Gould's great achievement was to demonstrate that the piece is also fun. He played with verve and swagger as well as intellectual rectitude, showing that the Goldberg Variations isn't merely a scholarly investigation of contrapuntal devices, it's a joyous piece of music. . . .His recording is rightly considered one of the landmarks of the LP era. In its wake, and thanks in no small degree to its example, the piece decisively entered the repertoire. Amazon currently lists 134 recordings!”

Finally, the kicker: Sony has recently issued, for all of $14.99, a 3-disc set consisting of beautifully remastered versions of both Gould’s original 1955 version of the Goldbergs and a remake he did in 1981, together with a third disc of interviews and outtakes. The versions are remarkably different, and critics are split as to which version is superior, but you don’t have to choose. I won’t go on about the Goldbergs or Gould performances -- anything I say would be mere surplusage to the countless extatic opinions expressed over the decades. Just read the reviews on Amazon or elsewhere and then run and get these discs.

• Gould's Remastered Goldbergs

What can I tell you? Different strokes . . .

What version(s) do you prefer?

• Casals Playing Bach on Naxos