article 5:   Notes on Bach's Variations for Harpsichord


Notes on Bach's Variations for Harpsichord
by Glen Wilson

Bach contributed masterpieces to most of the variation forms current in his day: the Passacaglia and the Canonic Variations for organ, the Chaconne for solo violin, the Crucifixus of the B-minor Mass ... even the Art of Fugue is a kind of massive variation-ricercar. Yet of variations for domestic keyboard - harpsichord or clavichord - he wrote only two sets; few enough, considering the great popularity of the form and Bach's striking devotion to the instruments. His first biographer, Johann Nikolaus Forkel, touches on the reason for this when he speaks of "variations, which, on account of the constant sameness of the fundamental harmony, he had hitherto considered as an ungrateful task."

This comment can only be understood in light of the fact that during Bach's lifetime, the usual variation form for harpsichord (aside from dance variations) was that which arose in Italy in the second half of the seventeenth century, known today variously as "thorough-bass", "bass-framework", or "formula" variations: the theme and all the variations have the same number of measures and the same fundamental harmonies. Everything else, including time-signatures and tempi, can be varied. The melody of the theme usually plays no further role, although composers north of the Alps who took over the form sometimes combined it with elements of older melodic-variation and cantus-firmus techniques.

Now, the main motor of Bach's music is the development of a limited number of melodic figures through highly dynamic harmonies. In the type of variation under discussion, the situation is reversed: the static harmony compels a search for ever-changing melodic formulas. To give each movement the motivic unity Baroque esthetics requires, each variation is usually based on one sole formula, hence the term "formula variations". And this system, says Forkel, Bach found too confining.

He wrote an early essay in this form at some point, the "Aria variata alla maniera italiana". It is included in this recording (note: this applies to the original Teldec release only) for a number of reasons: as the "other" set of Bach's harpsichord variations; to help put the Goldberg Variations in historical perspective; to show the distance Bach travelled in the twenty or thirty years between the two works; to illustrate changing tastes in harpsichord building and tuning; and, finally, for its own merit. For although the Aria variata is no great masterpiece, it is a charming example of his early style, and shows the master's hand in an extra attention to detail and a attempt to develop the "formula" in each variation that sets it apart from the routine products of the Italians of Pasquini's school, or from Handel's hollowly pretentious dance-variations. The lovely a-minor Aria still breathes the spirit of the old German style of Bach's mentors Reinken, Böhm, and Buxtehude, although the second half modulates with a boldness typical of the young Bach. He wisely saves his biggest surprise - a newfangled Neapolitan sixth chord - for the end of the last variation, which is basically the Aria da capo. Bach had an example for this procedure in a work by one of his cousins, and would return to it in the Goldberg Variations.

After an unfinished fragment in the first "Clavierbüchlein" for his wife Anna Magdalena (ca. 1722), Bach turned his back on variations for the harpsichord until twenty years later, when he would take them up again with a vengeance.

The well-known story of how he came to do so is in Forkel's above-mentioned biography. Modern musicology has cast doubt on the tale, for no good reason that I can see. Forkel knew and corresponded with Bach's sons C. P. Emmanuel and Wilhelm Friedemann, and had collected the material for his book by 1775, although it was only published in 1802. The atmosphere of a detailed family anecdote rings true. The passage deserves to be quoted in full, if for no other reason than that it has become so closely associated with the piece:

For this model, according to which all variations should be made, though, for reasons easily understood, not a single one has been made after it, we are indebted to Count Kaiserling, formerly Russian Ambassador at the Court of the Elector of Saxony, who frequently resided in Leipzig, and brought with him Goldberg, who has been mentioned above, to have him instructed by Bach in music. The Count was often sickly, and then had sleepless nights. At these times Goldberg, who lived in the house with him, had to pass the night in an adjoining room to play something to him when he could not sleep. The Count once said to Bach that he should like to have some clavier pieces for his Goldberg, which should be of such a soft and somewhat lively character that he might be a little cheered up by them in his sleepless nights. Bach thought he could best fulfill this wish by variations, which, on account of the constant sameness of the fundamental harmony, he had hitherto considered as an ungrateful task. But as at this time all his works were models of art, these variations also became such under his hand. This is, indeed, the only model of the kind that he has left us. The Count thereafter called them nothing but "his variations". He was never weary of hearing them; and for a long time, when the sleepless nights came, he used to say: "Dear Goldberg, do play me one of my variations." Bach was, perhaps, never so well rewarded for any work as for this: the Count made him a present of a golden goblet, filled with a hundred Louis d'ors. But their worth as a work of art would not have been paid if the present had been a thousand times as great. It must be observed that, in the engraved copies of these variations, there are some important errata, which the author has carefully corrected in his own copy.

Note that the Variations were meant to cheer the listener up, not put him to sleep, as is sometimes thought (and as sometimes happens). Interpreters should also note the words "of a soft and somewhat lively character".

Johann Gottlieb Goldberg, who was about fourteen when the Variations were composed, had been brought by Keyserlingk (Forkel modernizes the spelling) from Danzig at age ten to study with Friedemann Bach at Dresden. The eighteenth-century critic Reichardt writes: "They tell wonders of his incredible talent: he is said to have played the most difficult pieces with ease not only at first sight, but even with the score upside-down. The great J. S. Bach always spoke of him as his best student in harpsichord and organ. He occupied himself day and night with music and was not in the least interested in anything else."

It is fascinating to speculate on how Bach, then entering old age, would have reacted to such a child prodigy. The superficially most striking aspect of the Goldberg Variations - their technical brilliance and difficulty - possibly had one of its inspirations here. This, and his connection with the work as mentioned by Forkel, justifies to some extent the title "Goldberg" Variations, which has been in use for so long as to have established a kind of "eminent domain". I, for one, am attached to it, and prefer it to the original: "Aria with Diverse Variations for Double-Manual Harpsichord".

Another contemporary virtuoso who undoubtedly jolted Bach into an effort to emulate and surpass was Domenico Scarlatti. In the late 30's and early 40's of the eighteenth century, dozens of his Sonatas were published in London, Paris, and Amsterdam, and some of them must have come into the hands of Bach, who was always on the lookout for the latest in musical trends and even ran a kind of sales agency for music. Just as the publication of Handel's Suites in 1721 may have goaded Bach into issuing the Partitas, so the Thomaskantor was not the man to allow Scarlatti's quantum leap in idiomatic, virtuoso keyboard style to pass without a response. I would also mention, as pure speculation, another publication by an Italian: G. B. ("Padre") Martini's harpsichord Sonatas, superbly engraved at Amsterdam sometime around 1740. The date usually given is 1742, the year the Goldberg Variations are thought to have been published. But neither work can be dated with certainty, and it is also possible Bach knew the Sonatas in manuscript. In any case Martini's print contains two sets of Aria and Variations, a Corrente in Canon, and figuration and ornamentation more similar to those in Bach's work than any other I know of. Martini's whole language is more baroque than that of the "galant keyboard player" (as Quantz calls him) Scarlatti. Strangely enough, Bach's youngest son Johann Christian later studied counterpoint with Martini, who was to Italy what J. S. Bach was to Germany: the rear guard of the retreating polyphonic tradition.

Italy, then, having laid the groundwork for late baroque variations, also founded the school of keyboard virtuosity, which through Bach received one of its first expressions in Germany. And yet, of the influences pervading the Goldberg Variations, that of Italy's rival for cultural hegemony in Europe - France - is at least as strong, if less apparent. It is impossible to overestimate the ascendancy exercised by French culture over the German mind in this period, the waning of the Grand Siècle and the dawn of the Enlightenment. Voltaire, on his way to Frederick the Great in 1750, wrote, "I find myself in France. German is spoken here only by soldiers and mule-drivers." Bach, addressing the Margrave of Brandenburg - one of the hundreds of German aristocrats who wanted their own little Versailles - in this request for a little patronage in return for six immortal concertos, used French. Bach's own house, the Thomasschule, was rebuilt in 1732 with Mansarde roof and risalite, to look like a French Hôtel, as was much of the rest of Leipzig.

For that city, the circle around the poet, critic and professor Gottsched was especially important. He was said to have turned Leipzig into "a little Paris". This priest of French Classicism and prophet of the Enlightenment almost singlehandedly reformed German literature and criticism. His admiration for Bach is well-documented, his gifted wife studied harpsichord with Bach's pupil Krebs, and Bach himself set at least three texts by Gottsched to music, including the "Trauerode".

Gottsched's main concern was for order, unity, and "rational" rules in the arts, as a reflection of an ordered and rational view of the universe. The classic expression of this outlook is the French formal garden, in which nature is wedded to numbers and geometry. Across the street from the Thomasschule, behind the home of the patrician Bose family (with whom the Bachs were on godparent terms) was the most famous private garden in Leipzig - in the French style. From his workroom in the Thomasschule, Bach could look out across the river Pleisse and see a newly laid-out park with the same geometrical patterns.

These influences from his closest surroundings must have suggested to Bach the well-known ground-plan of the Goldberg Variations. Its numerical rigor and strict symmetries, unparalleled in the history of music, have always been much emphasized; let us review its main points. The thirty Variations are flanked by the Aria and the Aria da capo, like a long facade with pavilions at each end, making a total of 32 movements. The central axis is marked by Variation 16, a French (!) Overture which opens the second half of the work. All the variations which are multiples of 3 (except no. 30) are canonic, with two upper voices in strict imitation and a free bass line. Variation 3 is a canon at the unison, Variation 6 a canon at the interval of a second, and so on, the interval of imitation increasing by step until at Variation 27 (3x3x3), a canon at the ninth (3x3), the accompanying bass voice drops away. Moreover, the Aria and each Variation is a microcosm of the whole, having 32 measures in two sections of sixteen bars, both to be repeated. In a work of such well-planned proportions, I personally feel there is no alternative to taking all the repeats, and I would like to express my thanks to Teldec for respecting my wishes in this matter, even though it meant spilling over onto two CD's.

The plan of the Goldberg Variations is certainly masterly. Yet one could write a set of variations on the same plan and still have it come out as rubbish. Or one could play the Goldberg Variations, plan and all, and still play it badly. The plan, as I see it, gets too much attention; it is to the actual music as a menu is to a great meal, or a map to an exciting journey - a mere starting point. The music holds riches of which such poor "analysis" can give no inkling.

French influence did not stop at the planning stage. Bach took a cue from France for one of the unique aspects of the Goldberg Variations, hinted at in the original title. It is one of the rare works in the literature for harpsichord which really requires two manuals, and of those, it certainly exploits the opportunities presented to the fullest. In variations marked "a 2 Clav.", each hand passes freely up and down its own keyboard; they cross sometimes to extremes that must be seen and not just heard. The two voices intermingle, chase each other and leap about in the most delightful display of crowd-pleasing acrobatics and sound effects Bach ever permitted himself. French composers, however, had been writing for the harpsichord with two independent manuals practically since its invention around 1640, Louis or Charles Couperin having composed at least two such "pièces croisées". Among others, François Couperin (with whom Bach corresponded) made brilliant use of the idea. What Bach did was to couple the procedure with Italianate virtuosity and stamp it with his incomparable genius.

These particular Variations can only be played on two solo eight-foot registers. The mechanics of the historical harpsichord allow no other choice. The delicacy, lightness, and precision that this registration imposes are very French indeed. As exuberantly Italianate as the figuration seems, it has been refined and tempered into something my teacher Gustav Leonhardt once compared to "Regence filigree". Nothing could be further from the viciously motoric tension which is all most modern approaches have to offer in the more brilliant Variations. They are "soft and somewhat lively" - Meissen porcelain, not airline plastic.

As a final testimony of Bach's indebtedness to French music, there is the Aria itself, this coquette masquerading now as modest sweetness, now as bold flirtation, finally as languor, and then back again, her charms clothed in a glittering brocade of the richest French ornamentation; a conversation galante straight out of a novel by Madame de La Fayette.

A persistent misconception has it that the Aria was originally a Sarabande in the second Clavierbüchlein for Anna Magdalena, some going so far as to state it is so titled in that source. Actually, it is an untitled late entry in the book, made around the time of the composition of the Variations, and undoubtedly in connection with them. And if we must insist on giving it a dance-label (why can it not just be an Aria?), the Polonaise would be as good a choice. This dance had an extraordinary vogue in the mid-eighteenth century, appearing in a variety of guises, including the slow and graceful. Princess Anna Maria of Saxony, daughter of Bach's patron Augustus III, Elector of Saxony and King of Poland, had a collection of some 300 of them, Wilhelm Friedemann Bach wrote a set of twelve very beautiful Polonaises, and none other than J.G. Goldberg composed a cycle in all 24 keys, a kind of Well-Tempered Clavier alla polacca. Furthermore, Count Keyserlingk came from Kurland on the Baltic, a province long under Polish control. He would have appreciated the choice of a Polonaise as the theme for 'his' variations.

All these imported characteristics alone would not have served to remove a certain stigma of frivolity from the variation form for a German composer trained in the traditional school of counterpoint and imitative polyphony. By 1740 Bach had lost any interest he might ever have had in composing lightweight, pleasant music for a wider public. He was well aware of his powers and saw it as a God-given duty to write the best music he could - and to a German, that meant well-crafted counterpoint. To balance French order and grace, Italian virtuosity and exuberance, and German solidity, he hit upon the idea of the series of nine canons. Canon is by far the most difficult procedure in music, and to the "rule" (the literal meaning of the word "Canon") of exact imitation of one voice by another entering later, Bach here bound himself as well to the "rules" of the Variation harmonies and the stepwise increasing intervals of imitation. These appalling difficulties are offset only by the addition of the third, non-canonic bass voice. The canons at the fourth and fifth are in contrary motion: where the leading voice ("dux") moves up the follower ("comes") moves down by the same interval, and vice versa. They herald the approach of the work's centerpiece, the Overture.

The nine canons are an ingenious new "formula" for variations, which serve to increase the cycle's specific gravity, as well as articulating the long facade. They lie, as it were, in wait after each of the brilliant double-keyboard variations - reminders of what real compositional craftsmanship is. I suspect Forkel was thinking of them when he wrote, "Through the arts of counterpoint he learned to make the most intricate canons, at all intervals and in all classes of motion, so easy and flowing, that none of the art applied in their construction was apparent, such that they seemed much more like free compositions." This is good advice - do not be too self-conscious about the canonic aspect of these variations; they are to be enjoyed as the pleasing music they are.

With the virtuoso variations and the canons, Bach had "formulas" for two thirds of his plan. For five of the others (1, 2, 13, 19, 25) he wrote movements in the style of instrumental solo- or trio-sonatas, thereby showing his awareness of the direction keyboard music was moving. Two of these are gorgeously ornamented Adagios, one in the minor key (25), which Wanda Landowska called "the black pearl of the set". Three more variations (4, 10, 22) are various styles of fugue. In variation 7, "al tempo di Giga", dance music takes a bow, and the orchestral Overture (16) has already been mentioned. The last of the canons is followed by two contrasting Toccatas (28, 29), keyboard music pur sang. The first is feathery light, all a-whir with trills, the second a glittering Jubilate of trumpets and timpani.

After all this buildup, the coming final variation ought to be something special, the listener feels. It is, too, though perhaps not quite in the way anyone would expect. It is a riddle, a smile, a very wise philosophical statement, and a strangely touching invitation to the Aria to return and complete the circle. Variation 30 is a four-voice "Quodlibet" ("what you will"), a potpourri in which practically every note is derived from two folksongs. Their texts apply symbolically to the music just heard. In the first, the Aria speaks to us, explaining its long absence:

"Cabbage and turnips have driven me away, if my mother had cooked meat I might have stayed longer."

In the second, we reply:

"It's been so long since I was with you, come closer!"

This is Bach's dedication of the Variations to Man, who has been defined as "an animal that laughs".

Having prepared these Variations "for the refreshment of the spirit" (as the title page says), Bach had them printed. Ironically, 50 years later Leipzig had become one of the music publishing centers of the world, but Bach had to send the Goldberg Variations to Nuremberg. He probably lost money on the venture, the engraving and printing of music being, at that time and place, inordinately expensive. It was for reasons of prestige and professional honor that Bach took the trouble to add the work to his series of Klavierübung (Keyboard Exercise), in which he had to set out to give the world examples at the highest level of the major keyboard forms: the dance-suite in Part I (Partitas), orchestral transcription styles in Part II (Italian Concerto and French Overture), the Prelude and Fugue and Chorale Prelude for Organ and the Bicinium for domestic keyboard in Part III, and now, in Part IV, variations.

In an anonymous retrospective review of Bach's keyboard music published some years after his death, we read: "What riches in Bach's printed Aria and Variations for double-manual harpsichord! What demands are here placed on the player's technique and the art of expressive performance!" And further on, this summing-up: "Was Bach not the creator of an entirely new way of handling keyboard instruments? did he not give them the greatest melody, expressivity, and song in his playing? He, who possessed the profoundest knowledge of the contrapuntal arts (and artifices as well), knew how to make them subject to Beauty."

The Goldberg Variations mark the turning point to the last phase of Bach's work. His later publications (Canonic Variations, Musical Offering, Art of Fugue) all dealt with fugue and canon, with the keyboard again predominating. (As everyone should know by now, the Art of Fugue is for harpsichord or clavichord. This was proven by Gustav Leonhardt in 1952.) The full importance of that turning point only became apparent in 1975, with a discovery which confirmed the last line of Forkel's story quoted near the beginning of these notes. Bach's own copy of the original edition came to light in Strasbourg. Since his manuscript has not survived, this is now the primary source for the Goldberg Variations. It contains important corrections, and many added ornaments, but the real surprise was on the back page. Here Bach had notated in puzzle form, and roughly in order of increasing difficulty, "various canons on the first eight bass-notes of the preceding Aria". They number 14 (the numerological symbol for B + A + C + H), and their discovery more than doubled the number of his known canons. The first four use only the eight-note theme, in the others counterpoints appear in which the theme is incorporated.

As explained at the outset, the Bass is the real theme of the Goldberg Variations. The extent to which Bach exploits the canonic potential of this fragment simply passes the bounds of sober belief; and yet, a casual "Etc." at the bottom of the page shows that Bach thought even more could be done. The theme is, in fact, an old chaconne/passacaglia bass (often used in minor key, as for example in the "Lament" from Purcell's Dido and Aeneas, where the descending fourth is chromatically filled up). It is impossible to know whether it was before or after he had noticed its canonic possibilities, that Bach decided to use it for variations. I personally find it hard to believe it was sheer coincidence that the Variations were grafted onto such a fertile Bass, and suspect that Bach had a canonic sequel in mind from the beginning, like the group of four canons at the end of the Art of Fugue. The theme is also closely related to the Christmas chorale "Vom Himmel hoch", theme of the later Canonic Variations for organ.

Bach clearly set great store by these canons, In 1747 he entered one in the liber amicorum of an acquaintance (no. 11), calling it "canone doppio sopr'il soggetto" (double canon on the theme). Until 1975, nobody noticed that the "soggetto" was the first eight notes from the Goldberg Variation bass line. Number 13 appears in the only authenticated portrait of Bach. The amazing Thuringian peers out at us with a humorously challenging expression, asking us to try our luck at solving the riddle of a canon triplex a 6 Voc. He even had it separately printed - one tiny riddle-canon on a scrap of paper. His pride was justified, for no other music ever written so resembles the harmonia mundi, the music of the spheres: the theme, its midsection, and its falling fourth plus inversion, wheels within wheels and mirrors reflecting mirrors into infinity.

But with these canons we enter a realm where words are worse than useless. If the Goldberg Variations are cabbage and turnips, the 14 Canons are nectar and ambrosia. This is about as close to Olympus as man will ever get.

- back -