The "Goldberg" Variations

(BWV 988)

The "Goldberg" Variations is the last of a series of keyboard music Bach published under the title of Clavierübung, and is often regarded as the most serious and ambitious composition ever written for harpsichord. Based on a single ground bass theme, the variations display not only Bach's exceptional knowledge of diverse styles of music of the day but also his exquisite performing techniques. Being also the largest of all clavier pieces published during the Baroque period, the work soars high above others in terms of its encyclopaedic character. From this, it is often considered that it sums up the entire history of Baroque variation, the Diabelli Variations by Beethoven being the Classical counterpart. However, doomed perhaps by its requirements of virtuoso techniques from a performer, it was not as popularly known as the Well-Tempered Clavier, which was not even published during the composer's lifetime. Nonetheless, the work has long been regarded as the most important set of variations composed in the Baroque era: in 1774 Johann Philipp Kirnberger, one of Bach's pupils, referred to it as "the best variations", while in 1802 Johann Nicolaus Forkel, the author of the first ever biography of J. S. Bach, praised the work as the "model, according to which all variations should be made".

Clavierübung Series
The Origin of its Common Name
Historical Background and Authenticity
The Work's Character and Structure
The Hidden Aim of the Work?
Personal Copy of the "Goldberg" Variations

Clavierübung Series

The "Goldberg" Variations was published in 1741. To the first edition was given the following title:

consisting of
an Aria
with Diverse Variations
for the Harpsichord
with Two Manuals
Composed for Music Lovers,
to Refresh their Spirits, by
Johann Sebastian Bach:
Royal Polish and Electoral Saxon Composer,
Capellmeister, and Director Chori Musici in Leipzig.

Nürnberg: Published by
Balthasar Schmid."


The fact that published pieces represent only a minority of Bach's entire creative output should be understood in the actual historical context. He was an extremely busy man in his office, and the financial burden of publishing music in his time was far greater than we, from our 20th century perspective, might assume that it was. This being the case, Bach's decision to go ahead with the publication of Clavierübung tells us, to a certain extent, of his pride and ambition. This publication began with Partita, piece by piece, in 1726, three years after he was appointed as Thomas Cantor and the Director of Music in Leipzig. Both the scale and contents of the Partita suggest that Bach took its idea from his predecessor, Johann Kuhnau (1660~1722), who published Neue Clavierübung (1689 and 1692) during his term of office in Leipzig. In 1731 Bach put together the already-published six Partitas and reissued them as Opus 1, denoting his intention to continue with the series. He published Part 2, consisting of Italian Concerto and French Overture, in 1735, followed by Part 3, the so-called "German Organ Mass", in 1739. Up to this point every new publication appeared every four years, until the last of the series, "Goldberg" Variations, which appeared after an interim of two years. During this fifteen year period concerned with Clavierübung, no other pieces were published.

The title Clavierübung (Keyboard Practice) should not imply that the work is a study for beginners. Rather, this particular title appears to have been chosen, so that, under the generality of its scope, the various types and styles of music written for several different keyboard instruments can be accommodated. So did Bach exactly, producing not only various types of music but also those demanding high technical standards in performance. Bach focused his target audience to "Music Lovers" alone, as here we see no reference to those "desirous of learning" which we find in the title pages of his other works of educational intent, such as Inventions and Sinfonias and the Well-Tempered Clavier. In this way Bach was able to explore musical contents of higher dimensions. In another sense, this may also have caused an unfortunate consequence, in that the circulation of these works was not as high as anticipated, and many copies remained unsold, simply because they were too technically difficult for most of the middle-class amateurs who dominated the market.

Despite its high profile character, we know little about what Bach intended with the "Goldberg" Variations. For instance, there is no series number given to the title of the work, such has been given to the previous parts from one to three. So how should we interpret the lack of number "4" here? Christoph Wolff offers an answer from the point of view of the publisher's sales policy. The publisher, Balthasar Schmid of Nürnberg, was a well-known figure in the publishing business community of the day. Wolff points out that although Schmid had been involved in the preparation of the two immediately preceding Clavierübung parts, as one of the engravers, his responsibility for the "Goldberg" Variations was significantly increased from the previous project, for now he took over the management of the entire manufacture, printing and distribution of the new work. On this ground Wolff presumes that Schmid would not have wished to begin his association with Bach by printing just the concluding part of a series of works. Indeed his assumption seems reasonable if we imagine that Schmid would not want to receive further orders for the preceding parts of the series. There is also a problem, however, when we extend our examination to include Bach's usual stubborn character. Can we really imagine Bach conceding with the publisher's morose wish if the mention of "Part 4" were extremely significant for him? Thus it is also conceivable that Bach intended to separate the "Goldberg" Variations from the rest of the Clavierübung series, and decided not to call it "Part 4". Unfortunately we do not have any documentary evidence for either of these interpretations, and so we have to quench our intellectual thirst with imagination. There is some internal evidence in the pieces themselves, of course, on which we can rely to inform our judgement. One of the likely reasons is the number symbolism which lies at the heart of the work's conception. For example, Part 1 is written for a single-manual instrument; Part 2 is written for a two-manual instrument, and consists of two pieces written in two major national styles of the day; Part 3 is written for three manuals (or two manuals plus a foot pedal, to be precise), and is constructed around the number symbolism of "3", which defines the work's structure, let alone the number of subjects in the final fugue. In the "Goldberg" Variations, we cannot find number "4" with reasonable clarity at all. There seem to be some other specific purposes in the "Goldberg" Variations, which are radically different from the preceding parts of the series, although Wolff considers all the four parts "as a unified whole".

The Origin of its Common Name

The "Goldberg", the work's popular title, owes its existence to Forkel's famous account of the story first told in the beginning of the nineteenth-century. As far as we can trace, it was not the original title by the composer, at least at the time of publication. Nonetheless it is worth revisiting the facts, as Forkel received numerous pieces of first-hand, credible information from the eldest two sons, Wilhelm Friedemann and Carl Philipp Emanuel Bach. In commenting on the work, Forkel tells us the following story:

"For this model..., we are indebted to Count Keyserlingk, formerly Russian envoy to 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 fulfil 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."

There is much scepticism expressed about this famous episode. For one thing, there is neither documentary evidence of the work being commissioned, nor is there an official dedication in its published title-page, thus contradicting the custom of the day. For another, this celebrated Johann Gottlieb Goldberg (1727~56), whose skill on the harpsichord Bach knew, for he taught him when he was brought to Leipzig by Hermann Carl Reichsgraf von Keyserlingk (1696~1764) in 1737, would still have only been 14 years old when the work was published. Goldberg's compositions themselves do not display much of his acclaimed brilliance either. Lastly there is no mention of the golden goblet among the inventory of the estate when Bach died in 1750.

Supposing Forkel's account contains a certain amount of truth and we are to re-examine the affair once more, it is possible to speculate, at least, that Bach might have presented to Keyserlingk a manuscript copy of the work containing the dedicatory inscription. Or it may also be explained by an episode which emerged through the natural growth of their friendship since Bach presented to him a printed copy of the work during his Dresden visit in November 1741. Anyhow Bach must have felt grateful to Keyserlingk for his support in obtaining the title of Royal Polish and Electoral Saxon Court Composer in 1736. It would be quite natural for Bach to present him a copy hot from the press regardless of the work being commissioned. Even if the tale of the commission was true, it seems very unlikely that Keyserlingk's wish was the sole decisive factor determining the style and structure of the work, for, naturally we suspect, he would be too excited by the music rather than feeling sleepy. From 1741 onwards, we can also infer that their friendship became closer than ever, as Keyserlingk's only son began his study at Leipzig University in that year. It may have been part of human nature to exchange friendly jokes as their relationship was being established: "a heap of gold = Gold (gold) Berg (mountain)" can be one of such delightful witticisms uttered on one of their social occasions. The tale of insomnia, if we are permitted to pursue our guess-work to this extent, may also be a forged story with no evil intention, simply to make the story more attractive. The problem here is that both Friedemann and Emanuel, who supplied the information to Forkel, were not in Leipzig to witness the real episode. This clearly puts them in a weak position with regard to this kind of information. It seems that no definitive answer can be drawn until new factual evidence emerges.

Historical Background and Authenticity

Bach's published scores do not bear the date of publication. In the past, we used to procure an approximate dating from the newspaper articles of the day as well as letters that survived. More accurate dating was achieved fairly recently by examining the surviving printed copies themselves, extracting the unique information transferred from copper plates and the traces of corrections and annotations therein, while extending more widely the research into other works that were manufactured by the same publisher. It was Gregory Butler who identified that the "Goldberg" Variations was issued during the 1741 Michaelmas fair. The year 1741 was the time when Book II of the Well-Tempered Clavier was in the final stage of compilation. These two great monumental works were composed, to some extent, side by side.

Why Bach chose the form of variation is not entirely clear either, even if we may take the point of the work as being commissioned by Keyserlingk. Unlike Corelli and Handel, who wrote impressive examples of variations, Bach showed little interest in this form: the only examples prior to the "Goldberg" Variations were all minor works, such as the early chorale partitas and Aria variata in A minor. Towards the end of the Baroque era, there was a tendency toward variations of simple character, aimed for pedagogical use. Forkel's remark about Bach's view of variations as "an ungrateful task" precisely follows this trend. Yet the fact that Bach went against this trend with the "Goldberg" Variations invites the ambitions of scholars aspiring to solve the mystery. In any case, we can see a clear continuity in the trail of Bach's compositional activities from the Well-Tempered Clavier to the Art of Fugue, where he combined the form of fugue with the single-motive structure exhibited in the "Goldberg" Variations.

In the complete absence of Bach's autograph manuscripts, it is difficult to talk about the genesis of the work. The Aria placed at the beginning of the "Goldberg" Variations, which is repeated at the end, is in fact found in Book II of Clavierbüchlein for Anna Magdalena Bach (1725), copied by herself. In this instance, the Aria bears neither the name of the composer nor the title of the piece. Thus it is possible that the author of the Aria is anonymous. This is suggested further by the fact that the ground bass melody is a traditional theme (the earlier part of it at least), many other examples of which can be found in the repertoire of the seventeenth-century. On these bases, some claim that Bach merely borrowed his sixteen-years old ideas, while others argue that Bach was not the author of this Aria. The scholars working from style analysis also argue to such an extreme extent as to regard the modulation scheme and ornamentation as un-Bachian. Recently this contentious issue became the centre of open dispute, as two scholars, Frederick Neumann, who proclaims un-Bachian theory (1985), and Robert Marshall, who supports Bach's authorship (1976/89), bitterly impugned each other. The likely winner to be declared is Marshall, for his evidence of Anna's handwriting, which is identified as dating from the 1740s, is much stronger and more credible than those of his contender. Indeed it can be proven to some extent that she copied the Aria from one of Bach's autograph manuscripts, possibly the one he had used for writing the Stichvorlage (the exemplar for the engraving).

This new understanding from source studies seems to have overpowered the previously assessed evidence presented by the style analysis. In his book of 1993, David Schulenberg points out that "the Aria is neither Italian nor French but specifically German galant in style, and certain details point directly to Bach, especially the beautiful broadening out of the rhythm into steadily flowing notes in the last phrase."

The Work's Character and Structure

Being composed for two-manual harpsichord, the work's most prominent feature is its abundant display of modern, fashionable expressive elements of the high Baroque with a hint of Classical idealism, not to mention its magnificent architecture and formal beauty.

All the thirty-two pieces are built upon the same thirty-two-note ground bass and its implied harmony (one per bar in the opening Aria), the rhythm of which is maintained throughout the work. The fact that the ground bass line is always decorated melodically, and never appears in what-might-be-called the "original form", becomes one of the main features of the variations. In some movements, the theme acquires different harmonic flavours, while in others it is transferred to the high pitch range by the hand-crossing texture.

The Aria is a sarabande, a binary dance movement with repeats, consisting of two parts of equal length of sixteen bars each. This symmetrical structure is a prevailing feature in all the variations. In contrast, individual variations equip themselves with a unique character by having different time-signatures, the associated duration of the harmonic rhythm and other melodic materials. The concept of symmetry is also reflected in the overall shape of the work, as the thirty-two pieces are grouped into two parts. The second part begins with No.16, which is a French Overture. This piece is strategically placed after No.15, a canon written in G minor, so that the musical impact of the overture is effectively amplified by the vivid contrast, as if it is a new beginning. Thus the concepts of "number" and symmetry become the core of the work's structure and order.

When we examine how the pieces are arranged in the work, we notice that the same Aria reappears at the end (da capo) and in between these outermost movements are located thirty variations. Among these variations, nine are strict canons placed at a regular distance of every three pieces. The first to appear is the canon in unison (No.3), and until the work reaches the final canon in ninth (No.27), all the canons are systematically arranged in the ascending order of the interval between two canonic parts. It may now become obvious that the canons are organised around number "3". Inspecting more closely, we also find that the variations are also grouped in threes, consisting of free variation, variation in duet (mainly toccata), and canon, and there are ten such groups in all. The last variation is not a canon in tenth, as might be expected, but an unusual piece entitled "Quodlibet". A quodlibet is a contrapuntal piece built upon several different melodies, and here we hear two folk tunes: "I long have been away from you, Come here, come here, come here" and "Cabbage and turnips have driven me away, Had my mother cooked meat, I'd have chosen to stay". They are traditionally understood to be Bach's humorous wit hinting at the return of the Aria to conclude the piece. Amongst other variations, there are Two-part Inventions, Fughetta, French Overture, Trio Sonata, and various dance pieces, and through to the climactic moments towards the end, we experience, with ever increasing degree, the stunning technical display of keyboard writing, such as swift and flamboyant running passages, dazzling hand-crossing movements, and trills in inner parts, as if the composer is deliberately demonstrating his virtuosity in performance. The fact that Domenico Scarlatti published the acclaimed Essercizi in 1738, which Bach could have known, may be somehow related to this particular feature of the pieces.

The Hidden Aim of the Work?

The music analyses that explore various possible interpretations of the work, such as multi-layered structural divisions and groupings, may be intellectually satisfying. We cannot always determine, however, the extent to which they are based on Bach's real intentions (which can be justified from historical context) from hypothetical theories based primarily on our subjective invention. It is, indeed, very difficult to sift valid and authentic interpretations from so many speculative approaches which Bach's music attracts. For instance, in his article of 1984, David Humphreys claims that the unifying device of the work is the allegorical scheme of the work, which represents an ascent through the nine spheres of Ptolemaic cosmology, and he discredits the general consensus among listeners and players, that the work represents a purely musical unifying device, as a mistake. Humphreys first divides twenty-seven variations into three cycles, and names them as Canon (3, 6, 9 ... 27), Planet (4, 7, 10 ... 28) and Virtuoso (5, 8, 11 ... 29), then makes various attempts to associate Plato's cosmological philosophy and geometry with Bach's understanding of Affektenlehre seen in his means of expression. Still we are not at all sure whether Bach, who must have been deeply engrossed with his other duties, ever had the intention or inclination to undertake such highly abstract and profound mathematical stuff. It is amazing, to say the least, that the "Goldberg" Variations can also "refresh ... spirits" of musicologists who were not able to share such experiences from listening to the work.

There is another scholarly attempt which looks into Bach's ultimate aims of the work that deserves a special mention. In his article of 1987, Alan Street claims that Bach had a specific extra-musical intention, that is, his musical rebuttal of the attacks made on his compositional styles in 1737~38 by Johann Adolph Scheibe (1708~76). The most fascinating part of the scenario is that Bach, who was denounced by Scheibe for his lack of both general academic knowledge and the "true basis of music and its real beauty", in Scheibe's terms, decided to refute Scheibe by means of tactical employment of his knowledge of rhetoric in music of the latest style. Street argues that Bach found his resources in Quintilian's Institutio oratoria, a famous treatise by the ancient Roman orator that was popularly read in Bach's time. Based on the hypothesis that Bach used Quintilian's account of the orator's duty and resource as his ultimate inspiration, Street unfolds Bach's plan of reproach by taking Quintilian's description of forensic oratory, to address and rebut the charges. He claims that the technique of variation fits Quintilian's proposition that "best words are essentially suggested by the subject matter", interpreting that the 'subject of our speech' provides the 'best words' to be recast and reinterpreted in each variation. Examining more closely, he recognises that the Aria, acting as the exordium, is employed as that "most attractive form which draws its material from the speech [in this case, the professed taste] of our opponent", and that while Bach is being criticised by Scheibe in the first half of the variations, he refutes Scheibe in the second half. At "Quodlibet", the climax of the work, Street interprets "Bach's jest against things intermediate - cabbages and turnips are the stuff of the 'Goldberg' - and himself as 'mother'". If Street is correct, then we can also successfully explain some mysteries, such as why Bach did not call the "Goldberg" Variations "Part 4" of Clavierübung, and why he used the form of variations at this stage of his life.

Personal Copy of the "Goldberg" Variations

The above quoted Forkel commentary of the "Goldberg" Variations in fact continues as follows:

"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 copy."

This is confirmed in Bach's personal copy, discovered in Strasbourg in 1974, in which we can identify many corrections and additions carefully entered by him. Using red ink for most of the corrections, Bach perhaps intended at the time to publish a second revised edition, but which never materialised. Many of Bach's alterations remove engraving mistakes; yet among them we also find Bach's later improvements, which are the very significant part of the discovery. They include tempo indications to No.7 "al tempo di Giga" and to No.25 "adagio", performance-related marks, such as staccatos and slurs, and ornaments such as mordents and appoggiaturas, all of which is very useful information for performers. This is also the most important research material of the "Goldberg" Variations, quite apart from the manuscript copy of the Aria in the hand of Anna Magdalena, left for scholars to establish the process of revisions.

Among the revisions, the most significant finding must be the added appendix consisting of a previously unknown set of fourteen canons (BWV 1087), which is written very neatly in an unused page at the back of the volume. These canons are all based on the first eight ground bass notes of the Aria. Being arranged in the order of increasing contrapuntal complexity, these canons are set to include nearly all types of canonic techniques. Such thoroughness of Bach's technical display in the canonic writing can hardly be appreciable in the musical sense but only in the theoretical sense. It is normally believed that this number "fourteen" is intended by Bach to be his numerical signature (BACH = 2 + 1 + 3 + 8 = 14). In June 1747 he became the fourteenth member of the Society of the Musical Science organised by Lorenz Christoph Mizler (1711~78). To join the membership, Bach submitted the thirteenth canon found in the appendix of the "Goldberg" (albeit a slightly revised version, BWV 1076). This is the canon which appears in the famous portrait painted by Elias Gottlieb Haussmann (1746 and 1748). As it happened, Bach concluded the "Goldberg" Variations with his musical signature.

(c) Yo Tomita, 1997

This essay is inspired by the performance of the work on the harpsichord by Masaaki Suzuki, 1997, and is used for his CD released in the autumn from BIS.