It is well known that Bach aggressively
studied the music of his contemporaries and predecessors as he developed his own personal and unique style. In particular, his work transcribing Vivaldi’s string concertos is often cited as a watershed in Bach’s education. However, a closer look at the concerto transcriptions and their genesis will encourage us to re-evaluate their role in Bach’s stylistic development.
The transcriptions stem from Bach’s Weimar years, probably between 1713 and 1717. It is believed that much of the source material was provided by his patron, Prince Johann Ernst. In 1713, Ernst visited Amsterdam and purchased a large quantity of music, likely including Vivaldi’s newly published Opus 3, L’estro armonico.1 The chart to the right shows the extant concerto transcriptions made by Bach; there are 23 transcriptions from 21 originals.2 Bach was not alone in making concerto transcriptions; from Johann Gottfried Walther, his colleague in Weimar, we have 14 surviving transcriptions.3
The purpose of Bach’s concerto transcriptions has been debated and probed at length. At first, scholars were inclined to believe the words of Johann Nikolaus Forkel, who wrote in 1802 that Bach undertook the transcriptions for the purpose of education.4 However, the extent of Bach’s activity in this area seems to exceed the needs of self-improvement; one does not need to make dozens of idiomatic keyboard arrangements of concertos to learn how to write one for strings. And of course, if the purpose of the exercise were purely educational, there would have been no need to transcribe the works of the teenage Prince Johann, who was himself a student of Walther and Bach. Therefore, it is now widely believed that the transcriptions were actually commissioned by the prince, a theory first advanced by Hans-Joachim Schulze.5
Also difficult to discern is what Bach actually learned from Vivaldi. Forkel wrote that from Vivaldi, Bach learned “musical thinking” and the concepts of “order, continuity, and proportion.”6 As Christoph Wolff has asserted, this statement may be reliable precisely, and ironically, because Forkel had no knowledge of Vivaldi’s music and no way to know what Bach learned from it; therefore, the statement could well originate from Bach’s sons who were in contact with Forkel in the late 18th century.7 Nevertheless, there were many other Italian models at Bach’s disposal, not to mention the works of Telemann, an established master who was close at hand. And it has been observed that Bach was able to create a coherent ritornello form as early as 1708, in the opening movement of Cantata 196.8 Taking all that into account, perhaps it is more interesting to observe what Bach did not learn from Vivaldi: that is, what musical elements did he alter in Vivaldi and subsequently avoid in his own works?
The concertos Bach transcribed from Vivaldi’s Op. 3 provide the best avenue for this study. These works are the most elaborate of Bach’s transcriptions, and they were based on outstanding originals available to Bach in an authoritative published edition. His other Vivaldi transcriptions were made from manuscript sources of varying integrity.9
Op. 3 was Vivaldi’s first publication of orchestral music, an ambitious offering with the brazen title L’estro armonico, “harmonic inspiration.” Vivaldi chose the Amsterdam publisher Etienne Roger for this collection for two reasons: the superiority of Roger’s work and the opportunity to exploit the strong demand for Italian music in Northern Europe. Initial publication in Amsterdam was in 1711; soon thereafter, it was published by Walsh in London (1715 and 1717). Several French editions followed, beginning in the 1730s. Roger reissued the collection no less than twenty times, finally ending production in 1743. Its popularity only rivaled by Corelli’s Op. 6, L’estro armonico established Vivaldi’s reputation throughout Europe.
The publication was exceptional in that it consisted of eight part books: four violin parts, two viola parts, one cello part, and one part for double bass, which included the figures. A more typical concerto publication would be in just five parts, the solo part plus the usual quartet of string parts. In fact, Vivaldi’s later concerto publications were generally à cinque; none are in eight parts. In all cases, production of a score was left to the purchaser.
The eight-part presentation of Op. 3 allowed for considerable variety in solo groups: there are concertos for one, two, or four soloists. In addition, the cello is often emancipated from the continuo and is able to join the soloists in virtuoso passagework. One player per part is sufficient to perform a concerto; solo and tutti contrasts are provided by the doubling in the part writing, not by the use of a large ensemble. The bass part is fully and carefully figured, even in Vivaldi’s frequent unison passages (Illustration 1).
The structure of Op. 3 is ingenious. There are twelve concertos in four groups of three: the first of each three is for solo violin, the second for two violins, and the third for four violins. Superimposed on this scheme is a tonal arrangement in pairs, alternating major and minor keys, with the last pair reversed to end in major. Unfortunately, Vivaldi’s elegant concept is violated by most modern editions10 and obscured by the commonly used Ryom catalogue.11
There is also an intriguing logic to Bach’s approach to the source material. From the twelve concertos of Op. 3, he arranged three solo violin concertos for keyboard without pedal, two double violin concertos for organ with two manuals and pedals, and one concerto for four violins is transcribed for four harpsichords and orchestra. Although there may have been more transcriptions made and subsequently lost, these six arrangements seem to comprise an orderly exploration of the original material.
The manualiter concertos are probably the most neglected works in this genre. Robert Marshall makes the case that the classification of these as harpsichord works in the Bach index, and in editions of the keyboard works, is arbitrary, and that they are equally likely to be organ works.12 Various factors support this theory: One, there was a tradition of composing organ pieces both pedaliter and manualiter, sometimes in complementary fashion, as we find in Clavierübung III. Two, performing concertos at the keyboard was especially fashionable on the organ; in fact, the practice may have been first popularized by an organist in Amsterdam, Jan Jacob de Graaf, whose proficiency performing concertos at the organ was praised by Mattheson.13 Three, Bach’s primary role at Weimar was organist, not harpsichordist. Four, the manualiter transcriptions were transposed and adapted to fit the range of the organs played by Bach in the Weimar region, which was four octaves, from C to c′′′. In general, there is a modern tendency to overlook the need for 18th-century musicians to play organ music without pedals; such pieces would have been attractive to gentlemen amateurs, ladies, and young people, as well as professional organists in smaller churches. While there is certainly no reason to exclude one instrument or the other, organists should be aware that the manualiter transcriptions contain some excellent material rarely heard on their instrument.
We can study many of the traits of the manualiter transcriptions by looking at BWV 978 (Example 1).14 The transposition by Bach to F major avoids the note d′′′, which is prevalent in the original. More interesting is Bach’s complete reworking of the bass line; the left hand does not wait for the opening theme to be stated, but enters early with a closely related countermelody. Throughout the manualiter transcriptions, Bach adds passagework in the left hand, leaving the treble mostly unchanged. In mm. 7–11, Vivaldi’s homophonic eighth-note accompaniment is replaced by broken-chord sixteenth-note figuration in the left hand (Example 2). Perhaps a better solution to this problem would be found by a later generation with the Alberti bass.
Another trend in the manualiter transcriptions is Bach’s avoidance of manual changes and dynamic contrast. Note that the original’s echo is gone and the added counterpoint makes a manual change impossible (Example 1, m. 3–4). Throughout the manualiter transcriptions there is no attempt to render solo and tutti contrast with manual changes. There are only occasional dynamic effects requiring two keyboards, and these are for echo gestures within the tutti ritornello, as in Op. 3, No. 12 and BWV 976, m. 2.
The two best-known concerto transcriptions are those for organ with two manuals and pedals, in A minor (BWV 593) and D minor (BWV 596); these are part of the standard concert repertoire for organists and are on a higher level of virtuosity and complexity than the manualiter concertos. In the organ transcriptions, two manuals are consistently and effectively used for dynamics, solo with accompaniment, and solo-tutti contrast. The manual changes are clearly notated and the voice leading and beaming designed to accommodate them. Despite this successful experience adapting Vivaldi’s dynamic effects to the organ, Bach almost universally avoided manual changes and dynamics in his own organ works, the exceptions being the Toccata in D Minor, BWV 538, and the Prelude in E-flat, BWV 552/1. In other organ works, even those that are concerto-inspired, no manual changes are indicated and the counterpoint makes changes awkward.15
Again in the organ transcriptions we see Bach’s tendency to fill in the rests and longer note values with continuous 16ths, perhaps with a bit more finesse than in the manualiter transcriptions. In Example 3, he not only filled in the rests in Vivaldi’s original but also created a quasi-imitative sequence. The challenging sixteenth-note pedal passages Bach added in BWV 593/3, mm. 59–63, lend further weight to the argument that the transcriptions were intended for virtuoso performance rather than theoretical study.
Mm. 51–54 in the first movement of BWV 593 are peculiar for their use of octaves where Vivaldi’s original is fully harmonized, a rare instance where Bach is less full in texture than his model.16 Another oddity is the indication “Organo pleno” in m. 51; most likely, this is a copyist’s error for “Oberwerk.” It does not signal a registration change, but simply a return to the main keyboard with its plenum.
Sometimes exceptional means are used to create a solo and accompaniment (Example 4). It is strange, and perhaps disappointing, that Bach never used this kind of multi-layered symphonic texture in his own organ works.
BWV 596 in D minor is the only keyboard concerto that survives in autograph (Illustration 2, on page 21). It was long thought to be a work of Wilhelm Friedemann Bach because of the inscription “di W. F. Bach” followed by “in manu mei Patris descript” (“written in the hand of my father”). However, in this case “di” means “of” or “owned by”; Wilhelm
Friedemann was claiming ownership of the manuscript, not authorship of the piece. As a result of this misunderstanding, BWV 596 is missing from the Bach Gesellschaft, Peters, and Widor-
Schweitzer editions of the organ works.
The D-minor concerto is perhaps the most interesting of all the Weimar era transcriptions, and if the survival of an autograph is any indication, it may have been Bach’s favorite as well. One remarkable characteristic of the original is Vivaldi’s rigorous and energetic fugue, which exhibits ingenious invertible counterpoint as well as solo/tutti contrast. Surely, this piece served as inspiration for Bach’s concerto movements that synthesize fugue and concerto (e.g., final movements of Brandenburg Concertos Nos. 4 and 5).
The beginning of the concerto has attracted considerable attention for Bach’s unusual registration instructions: Oberwerk Octave 4′, Brustpositiv Octave 4′, and Pedal Principal 8′.17 This registration was not an aesthetic choice, but was contrived for a purely practical reason, to avoid the d′′′ prevalent in the original. Since transposition of the concerto to C minor would have made the fugue, with its fast parallel thirds and sixths, very awkward to play, Bach used 4′ stops and played the opening section an octave lower. This registration should not be considered a model for registering other concerto movements and playing entire movements on a single principal stop. It is a unique exception to the normal registration for concerto fast movements, which is organo pleno.
Bach made an interesting change in this opening passage, rewriting the two solo violin lines to make a strict canon and adding an extra measure where the canon winds down (m. 10).18 This change is unique in Bach’s transcriptions; normally, he maintained the dimensions of the original, neither adding nor subtracting measures. The addition of a canon to this concerto conflicts with the traditional view, stated by Forkel, that Bach used Vivaldi as a guide away from improvised “finger music” toward a more intellectual and organized approach to composition. In this passage, BWV 596 is clearly more cerebral than the model.
At the end of the fugue, Bach made a significant change (Example 5); in order to effect a stronger conclusion, he added more harmonic interest, rhythmic drive, and a Picardy third ending.
Another interesting change is found at the end of the last movement (Example 6). Vivaldi’s tremolo string writing is fruitless on the organ, so Bach used sustained chords in conjunction with a newly added tenor line. The added line is sufficiently violinistic that few organists suspect it is not original to Vivaldi. Bach used nearly the same tenor figuration to replace a tremolo passage in another Vivaldi concerto; see Op. 7, Bk. 2, No. 5 and BWV 594/1, mm. 26–27, 32, 34, etc.
That these two movements were altered to end with a major chord is revealing. Such a change is unnecessary in the context of a transcription, and thus represents a purely aesthetic choice made by the arranger. Comparing how each composer ends minor key movements leads to some striking differences. In Op. 3, there are 24 minor-mode movements; none ends with a Picardy third. Further, in the Op. 4, 7, and 8 concertos one searches in vain for a Picardy ending. Bach did not publish any large sets of concertos; nevertheless, we can observe that all six Brandenburg concertos are in major keys—which may be significant in and of itself. Of the minor-key slow movements, only one ends on a minor chord. One ends Picardy and another two end with a Phrygian cadence, more in the manner of Corelli than Vivaldi. Looking at some other organized sets of Bach works from Weimar or soon thereafter, we see that in the Orgelbüchlein and Well-Tempered Clavier I every minor-key piece ends with a major chord, except one (BWV 863/2).
There are other significant tonal differences one could explore; Vivaldi often tends to have all three movements in the same key, and in some cases will have the slow movement of a minor-key concerto in the subdominant, also minor. On the other hand, Bach will more typically use a mediant relationship for the middle movement, exploiting the relative minor or major. Ending a major-key movement, Vivaldi will stay in tonic, without hint of other keys; Bach will usually tonicize the subdominant just before closing. All of this leads to the conclusion that Bach did not emulate Vivaldi in some crucial matters of harmony and tonality.
The last concerto Bach transcribed from Vivaldi’s Op. 3 was the Concerto for Four Harpsichords and Strings in A Minor, BWV 1065, based on concerto No. 10 for 4 violins in B minor. This transcription is much later than those for keyboard solo. Stemming from around 1730, it is a Leipzig work destined for performance by Bach’s Collegium Musicum. Here we find little trace of Bach the learner, as he takes a fine Vivaldi original and puts his own stamp of genius upon it, enriching the texture and harmony throughout. Of particular interest is the poignant chromaticism added to Vivaldi’s diatonic sequence in mm. 82–85, and the 32nd-note keyboard flourish in mm. 90–91, the latter similar to some passages in Brandenburg Concerto No. 5.
In conclusion, there can be no doubt that Bach learned certain elements of composition from working with Vivaldi’s models; indeed, Op. 3 was a musical landmark that influenced most composers in the early 18th century. However, there is sufficient musical evidence in the transcriptions to suggest that Bach was a mature, confident, and highly original composer in the early Weimar years, before he made the concerto arrangements.