The Rust Variant BWV 903a
1. Editor’s Praeambulum
The following essay on the Rust Variant BWV 903a is a thought-provoking article that, written with the conviction and rigor characteristic of Dr. Tureck’s personality, attempts to invalidate the authenticity of the Variant by providing a wealth of historical and analytical evidence finely thread throughout three extremely detailed sections. Its main hypothesis—that the Variant is not only unauthentic, but also the daring creation of Wilhelm Rust—is already established in its opening part and further developed in Parts II and III.
The essay presented here is the most recent and complete version of Dr. Tureck’s study of BWV 903a and corresponds to material typed in July 1991. In Part I (‘Source, Provenance, Acceptance’), Dr. Tureck surveys the reception of the Variant and summarizes how it was approached by important editors throughout the nineteenth and twentieth centuries. Part I also investigates the historical circumstances surrounding the lives and work of the Rust family, in whose estate the supposedly putative Bach manuscript originated. In the light of this research, and particularly of data concerning Wilhelm Rust’s corrupt editorial practices and the stylistic unfitness and compositional weakness of the Variant within the context of the entire Fantasia, Dr. Tureck advances her hypothesis that BWV 903a must be a passage actually composed by Rust rather than an earlier version of BWV 903.
Dismissing the idea that “progress to a better state necessarily and universally moves from a vacuous simplistic form to a highly charged complex one”, Dr. Tureck then proceeds, in Part II (‘Bach’s Early Composing Procedures in Relation to BWV 903a’) to a comparative analysis between the Rust Variant and the compositional techniques featured in various organ works, mono-figural preludes, toccatas and other pieces including two of the Brandenburg Concerti. This examination, which throws light onto Bach’s “inborn capacity for integrated structuring”, forms the backbone supporting Dr. Tureck’s criticism of the Variant—which she regards as an incongruous, weak passage with little, if any at all, relatedness to Bach’s creative mind. Part III (‘Comparative Structural Analysis of BWV 903 and BWV 903’), finally, is a comprehensive, measure-by-measure comparative account of the structural features of both BWV 903 and BWV 903a concluding that, even if the latter was a sketch, an early version by Bach, consideration of its reception and musical substance should reveal enough evidence as to approach the Variant with a great amount of caution and skepticism.
While Part I provides much historical information on the appearance, disappearance and reception of the Rust Variant, parts II and III are of analytical nature. Dr. Tureck’s analytic style is profusely descriptive and extremely oriented to detail; it relies heavily on verbal depiction of compositional features and procedures. The reader will not find here graphs in the Schenkerian manner or in any other of the perhaps more ‘scientific’ fashions favored by contemporary approaches to musical analysis. However, it is not difficult to acknowledge certain similarity between Dr. Tureck’s approach and some of the Schenkerian principles, particularly in its insistent attention to structure, revealed by analytic language such as “reduced to its innermost harmonic core, the Fantasia may be said to be based in a seventh and its resolution” (p. 83).
In essence, the comparative analysis employed by Dr. Tureck confronts BWV 903 and BWV 903a in a number of structural minutiae: BWV 903 is based on a principle of floridity, and is rich in variety of contours, inversions, diminutions, changes of direction (alternation of ascending and descending designs), polyrhythmic figurations, harmonic subtlety and sustained integrated relationships. By contrast, BWV 903a, in its adherence to a continuous triplet figure, is based on a repetitive, unvaried motive, thus resting on a compositional principle of duplication alien to the floridity prevalent in Bach’s music and, as a result, lacking structural linkage to the Fantasia as a whole.
Due to a number of practical reasons, it has not been possible to enhance the essay’s text with musical illustrations and examples. Reading Parts II and III therefore requires an attentive attitude, and their understanding will be greatly enhanced if any recent edition of BWV 903 and BWV 903a can possibly be available while reading in order to locate musical illustrations at a glance.
When considered within the framework of both past and recent scholarship on the Chromatic Fantasia and Fugue, it seems appropriate to acknowledge that Dr. Tureck’s vision of the Rust Variant, condensed in this essay, is subjective in nature and extraordinary in its conclusions. To our knowledge, no single edition of the work (including the recent and reputable Wiener Urtext and the Urtext of the Neue Ausgabe sämtlicher Werke (Bärenreiter), both from 1999), and no scholarly publication have expressed a view similar to Dr. Tureck’s. All academic research on, and all published editions of, the Chromatic Fantasia accept the BWV 903a as an early version of BWV 903. Moreover, many editions provide the note-text of the Rust Variant as a viable, performable first section of the Fantasia. For instance, in the prefatory ‘Notes on Interpretation’ of the Wiener Urtext Edition, Michael Behringer explicitly states that “the versions published in this edition should not only be considered as editorial groundwork but also made practical use of” (Behringer, ‘Notes on Interpretation’, Chromatische Fantasie: XII).
In the insightful article “‘This fantasia… never had its like’: on the enigma and chronology of Bach’s Chromatic Fantasia and Fugue in D Minor, BWV 903”, George B. Stauffer affirms that “the surviving copies point rather unambiguously to three stages of development” … in the first of which, “represented in the early variant BWV 903a found in Darmstadt Mus. MS 69 (and the now-lost J. L. A. Rust manuscript, whose text served as the basis for BWV 903a as printed in the Bach-Gesellschaft and elsewhere, the Fantasia had a different incipit.” (Stauffer, ‘Enigma’: 173)
Dr. Tureck’s unique opinion of the Rust Variant may cause surprise to performers, scholars and students alike. However, the posthumous appearance of this essay should, at least, reignite the debate on authenticity, hopefully promoting a more critical approach to the indiscriminate acceptance of this version of the Fantasia. Dr. Tureck, whose scholarly authority has been admired as well as criticized , manifested also a powerful dislike of the idea of an ‘intermediate’ version corresponding to the copies of Mus. MS P803, in Johann Tobias Krebs’s and Samuel Gottlieb Heder’s hands. In a letter to dated July 1996, Dr. Tureck’s states:
“Whether or not approved by scholars of posterity, those two measures in P803/122 do not justify categorizing this manuscript as an ‘intermediate’ stage. The forcing into neat cubbyholes of ‘early’, ‘intermediate’ and ‘final’ as assigned to Johann Sebastian Bach’s creative stages in this composition, does not rest on a strong enough foundation.” (Personal communication with Oxford University Press, July 31, 1996)
As I also indicate in my preface to the ‘Critical Notes on Text and Interpretation’, Dr. Tureck’s research finished before the 1999 Neue Ausgabe sämtlicher Werke of the Johann-Sebastian-Bach-Institut Göttingen and Bach-Archiv Leipzig (Uwe Wolf, editor), and before the discovery in the United States of a previously unknown manuscript fragment in the hand of Bach’s second youngest son, Johann Christoph Friedrich, copied in the late 1740s with the authentic Latin title ‘Fantasia chromatica.’  Although it is difficult to assess what version was transmitted in this manuscript, the music notated thereon seems to correspond to the main version contained in BWV 903, rather than to BWV 903a.
Whether or not Dr. Tureck’s essay on the Rust Variant arrives to a definitely true conclusion may be less important an issue as it is to acknowledge that, perhaps even more urgently today than throughout the years during which Dr. Tureck researched the mysteries of BWV 903 and BWV 903a, modern performers could highly benefit from critically questioning the many assumptions that prevail in the domain of musical interpretation. If ‘authentic’ and meaningful performance is to occur, it is fundamental to encourage a healthily inquiring attitude toward musical compositions of previous historical eras. Dr. Tureck’s study of the Rust Variant is no doubt a brave and commendable example of a legendary performer’s and a scholar’s commitment to this ever-evolving approach to the performance of Bach.
1.1. Editorial Principles and Acknowledgements
The original material revised and presented here is archived in the Rosalyn Tureck Collection at the Howard Gotlieb Archival Research Center, Boston University. The editor wishes to acknowledge his gratefulness for the assistance provided by its staff in the preparation of this project.
M., m., Mm., mm.: measure, measures.
Mus. ms., Mus. MS: manuscript.
Footnotes and endnotes: There are two sorts of notes in the following essay, footnotes and endnotes. Footnotes are flagged by a superscripted number and are always entered by the editor. Endnotes are signaled as bracketed roman numerals and are Dr. Tureck’s original notes.
Although the original 1991 essay is a very advanced version, I have occasionally deemed it necessary to add further information to Dr. Tureck’s original text. This is the case of the brief commentary on two recent editions of BWV 903 in Part I. References and quotations have been rechecked and corrected inasmuch as the sources have been available. In those cases where quotations lacked a matching reference to the original sources, alternative passages have been provided, and their sources, acknowledged. Furthermore, certain citations from old dictionaries have been replaced with the most up-to-date versions thereof, and they have been indexed accordingly.
Dr. Tureck’s original writing uses both British and American spelling. For the sake of consistency, all original essays have been edited in American English.
Xoán Elías Castiñeira Varela
2. Source, Provenance, Acceptance
A central issue in any exploration of the Chromatic Fantasia and Fugue is the authenticity of those measures that have come to be known as the Rust Variant, BWV 903a. The first appearance and comment on this material, mm. 3–24, appears in the Griepenkerl/Peters Edition, 1843/44, forty-one to forty-two years after the first printing of the work in 1802. The music of these particular measures was accepted at the outset as originating from Johann Sebastian Bach and became certified by its inclusion in the Bischoff/Steingräber Edition of 1880 and the Bach Gesellschaft/Naumann Edition of 1886. It has been universally agreed that, because of their simplicity, these measures are an early version of that contained in BWV 903. Following a search of primary sources in the reception of the throughout over a century and a half, and conscientious examination of the structure of BWV 903a and BWV 903, it is my profoundly considered opinion that the simplistic section is utterly incongruous in partnering the treatment of the rest of the Fantasia and further, that it must be seriously questioned as to whether or not it emanated from the brain of Bach and that its inclusion in the Fantasia, either in print or performance, represents a genuine disservice to the composer.
The basic harmonic outline of these measures is, indeed, clearly similar to what is regarded as the later version. However, when compared with Bach’s structuring in sections of his early compositions that contain an appearance of simplistic treatment, the cast of mind revealed in this examination of BWV 903a appears to be one that is incapable for the structuring native to that of Johann Sebastian even in his earliest years. It has been recognized by everyone for a century and a half that the Rust Variant did not possess the qualities of BWV 903, but its weakness has been excused on the assumption, virtually universally accepted, that it was an early version.
The Rust Variant thus constitutes an indispensable adjunct to the study of the Fantasia. Its interest lies in a passage of twenty-one measures—mm. 3–24. Although the basic harmonic outline of these measures is similar to what is regarded as the later version, all of the latter’s rhythmic variety is absent; the figuration appears in the Rust Variant as a single sixteenth-note triplet figure, repeated invariably until mm. 21–24 when an abrupt change to thirty-second notes takes place. Moreover, the subtlety of harmonic relationships and treatment of BWV 903 is non-existent, neither stated nor implied. The rest of the Fantasia in BWV 903a, following m. 24, contains only comparatively minor variants. What is striking, then, is the incongruous content and style of the passage in question, and its conjunction, side by side, with the brilliance of the diversified figurations, harmonic treatment and subtle structural devices present in the surface texture, as well as secreted within the passages in the rest of the Fantasia (1).
No extant manuscript associated with Bach’s circle shows this version, nor, with one exception, does it appear in other extant sources and their branches (2). The exception is a manuscript whose scribe, source and provenance are not known (3). The prestige of BWV 903a is founded not on this manuscript but on the now lost manuscript allegedly in the hand of Friedrich Wilhelm Rust, whose elder brother, J. L. A. Rust was, indeed, of Bach’s circle, and whose rather dim history shall presently be traced.
The Rust Variant has been judged to be an “early” version (4) due chiefly to the simple triplet figure and its unvarying repetitive treatment throughout eighteen of its twenty-one measures. This view has retained its grip upon all editors and authors who acknowledge the dissimilarity between the measures in question and the rest of the Fantasia. The disparity is accounted for by attributing it to a first, less developed stage. However, no hard, documentary evidence has emerged in support of this hypothesis.
A simpler version is not necessarily a first stage. Yet, the assumption has been made, almost universally, that BWV 903a is an earlier version. No consideration has been given as to whether Bach’s own judgment would admit coupling a glaringly discrepant version as the Variant’s mm. 3–21 with the rest of the Fantasia, which is so rich in its rhythmic figurations and so imaginative in its shifting relationships. Also, when coupled with the Rust Variant, the irrelation of mm. 3–21 to the density and subtlety in the largest portion of the Fantasia from m. 24 to the end at m. 79 has received neither comment nor questioning except for a complacent supposition that first thoughts are simple and complexity is the result of revision.
What is responsible for this assumption of the chronological priority of BWV 903a, and its general acceptance right up to, and including, the Henle edition of 1969/1970?  Demonstrably, it is the nineteenth-century positivist style of thought, which applies the operation of linear evolution not only to all manifestations of like as a blanket rule, from the simple to the complex, but to the creative process as well. In the twentieth century, this linear mode of thought has lost its credibility. The evaluation of simple/early, complex/late is based on an outdated position. It has been applied to BWV 903a through a hundred and fifty years. At the end of the twentieth century, a fresh analytic scrutiny is long overdue.
2.2. History of the Reception of the Rust Variant BWV903a
The history of the Rust Variant’s reception may be surveyed from the following chronological conspectus of its transmission and reception from the mid-nineteenth century through the late twentieth century. It will be seen that, throughout almost a hundred and thirty years of publications, the assumption of a nearly version is virtually unanimous.
2.2.1. Friedrich Konrad Griepenkerl, 1843/44 (5)
The first mention of the Rust Variant is made by F. K. Griepenkerl 25 years after his original edition of 1819.
The following opening paragraph of the Preface to the third issue, 1843/44, of the Griepenkerl/Peters Edition of the Chromatic Fantasia and Fugue, refers to a “manuscript in the possession of Rust”:
Nach den gedruckten Ausgaben von J. N. Forkel und F.K. Griepenkerl (letzere v. J. 1819), womit noch sieben andere, teils ältere, teils neuere Handschriften verglichen wurden. Die älteste Form, in Bezug auf die Fantasie, scheint sich in der Handschrift bei Rust, v. J. 1757, zu finden, welche im Anhange No. 1 als Variante mitgeteilt ist. Die (…) unter No. 2 mitgeteilte Variante ist das Ergebniss einer Abschrift aus dem Nachlassse von Krebs… (6)
“After the printed editions by J. N. Forkel and F. K. Griepenkerl (the latter in the year 1819), which have been compared with seven other manuscripts, some earlier, some more recent. In respect to the Fantasie the oldest form appears to be that of the manuscript in the possession of Rust, dating from the year 1757, which can be seen in the Appendix No. I containing the variant. The other, under No. 2, is a copy received from the estate of Krebs…” (Translation by Rosalyn Tureck)
The Rust Variant is contained in this issue. Note that Griepenkerl does not single out an specific Rust. He mentions solely “Handschrift bei Rust, v. J. 1757, zu finden”. His reference is apparently to the provenance of the manuscript. Ambiguity emerges from the absence of Christian names or initials. This Rust would have to be Wilhelm Rust, descended from the well-known musical Rust family, and later to become the “chief of the Bach-Gesellschaft”. Since at least three different Rusts are involved, the omission is significant: it adds to the difficulties that will be seen to arise in identifying the precise Rust (7).
2.2.2. Hans Bischoff, 1880 (8)
In 1880, Hans Bischoff cites the manuscript by referring to its printing in Peters editions. He mentions no provenance; no specific citing of a Rust is made. Bischoff does not include the Rust Variant in his edition.
… endlich die alte Rust’che Handschrift vom Jahre 1757 deren Abdruck bei Peters S. I, C.4 zu finden ist. (9)
“… (I have compared)… finally the old Rust manuscript from the year 1757, which is found printed in Peters S.I., C.4.”
Bischoff makes no further comment on the Rust manuscript; his reference, Peters S.I., C.4. (Peters, Series I, Cahier 4), is to the 1866/67 reprint of the 1843/44 Edition quoted above (10).
2.2.3. Philipp Spitta, 1880 (11)
Griepenkerl hat zu seiner Ausgabe der chromatischen Fantasie und Fuge (P.S.I, C.4 Nr 1) zwei Varianten gefügt, von denen die zweite keine selbständige Bedeutung hat, die erste aber, nach einer Handschrift des Dessauer Capellmeisters F.W. Rust vom Jahre 1757, das Werk in älterer, vermutlich ursprünglicher Fassung zeigt. (12)
“Griepenkerl has added to his editions of the Chromatic Fantasia and Fugue (P.S.I, C.4, No. 1-207) two variorum readings, the second of which has no independent value, but the first, which is derived from a manuscript of the Dessau Capellmeister F. W. Rust from the year 1757, shows an earlier and possibly original form of the work”. (Translation by Rosalyn Tureck)
Contrary to his generally pervasive zeal, Spitta relies on the printed edition alone and does not reach beyond it to check the manuscript itself, although he was writing at the time that Bischoff, Naumann and Wilhelm Rust were alive, and presumably the manuscript was available for perusal. The Rust Variant, then, escapes his direct personal examination. Note that Spitta does not mention a lost cover. However, he is the first to pinpoint a specific Rust—F. W. Rust.
The usage of “des” in Spitta’s attribution “nach einer Handschrift des Dessauer Capellmeisters F. W. Rust…” again conveys some ambiguity to modern interpretation and our contemporary need for finely focused meaning in language. Although “des” represents ownership or possession, it may also carry the connotation of F. W. Rust as scribe. If Spitta is implying that the manuscript is in the hand of F. W. Rust, then this is the first scribal assignment so far encountered. However, a precise, unequivocal interpretation seems unattainable.
2.2.4. Ernest Naumann, 1890 (13)
In his edition of 1886, Ernest Naumann refers to the cover as lost, adding that it contained the names of J. L. A. Rust and the date, 1757 (14). How did Naumann know, if the cover was lost, that it contained this signature and date? How and from whom did he receive this information? This is the first mention of a cover and of J. L. A. Rust in that connection. Had Naumann seen that cover? He offers no further comment as to that, and no information on the scribe, noting only that the manuscript is in the possession of “Herrn Prof. Dr. Rust”, who is, of course, Wilhelm Rust. Naumann, curiously enough, omits any mention of F. W. Rust.
Since this Variant was printed initially in a Peters edition well before that of the Bach Gesellschaft (15), the note-text was already available in printed form. Naumann includes the Rust Variant in his edition. Was this printing in volume 36 of the Bach Gesellschaft Edition taken from the original Rust manuscript, from the 1843/44 Peters Edition, or from the 1866/67 Peters edition?
The question as to Naumann’s source arises inevitably when he states that the Variant is in the possession of Dr. Rust, his active chief in the production of the Bach Gesellschaft Edition. The initial assumption is that Naumann would have had easy, unobstructed access to the Variant manuscript itself, in preference to consulting and reproducing a twice-printed source in his Urtext edition. But Naumann contributes no enlightenment to posterity on this matter.
“Alte Handschrift im Besitz des Hernn Prof. Dr. Rust, mit dem Titel: ‘Fantasie Chromatique pour le Clavecin del Sigre. J. S. Bach’. Auf dem verloren gegangenen Umschläge derselben hat ‘J. L. A. Rust, Bernburg, 1757’ gestanden. Die Fantasie ist ohne, die ‘Fuga Chromatica’ mit Vorzeichnung des b, di r. H. im Discantschlüssel geschrieben.” (16)
“Old manuscript in the possession of Hernn Prof. Dr. Rust, with the title: ‘Fantasie Chromatique for the harpsichord by J. S. Bach’. On the lost cover was written ‘J. L. A. Rust, Bernburg, 1757’. The Fantasia is without, the ‘Fuga Chromatica’ with, the key signature of b, the right hand is written in the soprano clef.” (Translation by Rosalyn Tureck)
2.2.5. Heinrich Schenker, 1910 (17)
Takt 3. In der sogennanten Rustschen Handschrift erscheinen die Takte 3–20 in anderer Fassung. Dieselbe findet man mitgeteilt in der Ausgabe B.G., Anhang I, Seite 219 und bei Pt., I. Es wird—vielleicht für immer—leider völlig unaufgeklärt bleiben müssen, welche Bewandtnis es damit eigentlich hat. Soviel ist jedenfalls klar; die Fassung bei Rt. Steht, da ihr alle Mannigfaltigkeit in den Figuren fehlt, eben vom rein künstlerischtechnischen Standpunkt aus betrachtet, tief unter der in den übrigen Handschriften und auch hier im Texte mitgeteilten Fassung. Es mag geboten sein, daraus zu schliessen, dass möglicherweise ja Bach selbst bei einer späteren Bearbeitung von der ursprünglich noch starren und monotone Haltung zur freieren und mannigfaltigeren Fortschritt—ähnlich wie z.B. Beethoven, freilich in grösserem Umfang und Stil von der Leonoren-Ouverture Nr. 2 sich zur technisch verbesserten Ouverture Nr. 3 erhob—, da es doch kaum anzunehmen ist, dass sonst angesichts der schönen Vollendung der im Texte mitgeteilten Fassung es gerade den Verfertiger der Rt.-Handschrift gelüstet hätte, sich noch auf eigene Rechnung in einer minderwertigen Gangart der Motive zu versuchen. Wer wollte indessen der Einsichtslosigkeit so mancher Abschreiber und Herausgeber eine Grenze setzen! (18)
“Measure 3. Mm. 3–20 appear in the so-called Rust Manuscript in a variant. This Variant is also to be found in the B.G. Edition, Appendix I, p. 219 and in Peters, Variant I. It must unfortunately remain completely unclear, perhaps forever, what the situation here actually is. This much is clear, at any rate: the Rust Variant, lacking as it does in variety in the figures, is inferior from a purely artistic-technical viewpoint to the Variant given in the other manuscripts as well as in this text. It may be proper to conclude, therefore, that possibly Bach himself, on the occasion of a later reworking, proceeded from the originally stiff and monotonous attitude to a freer and more varied one—as for example, did Beethoven, to a greater extent and style, when he went from the Leonore Overture No. 2 to the technically improved Overture No. 3—since one can scarcely assume, confronted with the beautiful perfection of the Variant in the text, that the scribe of the Rust Ms. would have wanted to try his hand himself at an inferior structure of the motifs. Indeed, who would want to set a limit to the lack of understanding of some copyists and editors.” (Translation by Rosalyn Tureck)
Schenker voices the opinion that the situation of the Rust manuscript remains “unclear, perhaps forever”. However, he is still bound to the notion that the simpler version must be an earlier one, that “it may be proper to conclude” that the more complex one was the harvest of a “later reworking”. Schenker’s broad statement comparing Beethoven’s composing process with that of Johann Sebastian is rather wide of the mark. In these two minds and temperaments, so differently attuned in form, structure and cultural roots this process was not equivalent. We know that although at times Bach did revise, as in the Cantatas (19), there is no likeness to Beethoven’s composing process involving revisions. Bach’s revising procedures are not comparable to those multiple stages of labor found in Beethoven’s sketchbooks. These written out stages, left unpremeditatedly to posterity, represent the more modern—that is, the nineteenth-century sense of labor and progress, struggle and resolution, expressed, perhaps unconsciously, in his working out processes. On another level, these multiple steps are tantamount to Beethoven’s thinking aloud (20). Moreover, the sketches do not necessarily show a consistent growth in complexity. On the contrary, simplicity sometimes was Beethoven’s final choice (21).
Schenker is the only editor who adds repeat indications with a double bar to the Rust Variant (22). There is not the slightest indication, historically, formally, or structurally, that these 21 measures call for a repetition. It is remarkable that Schenker, who so rails against the “liberties” and “lack of understanding” of others, throughout the encompassing thirty-eight pages in his edition (23), should himself presume to insert repeat marks. Furthermore, he offers no explanation for these extraordinary additions to the Rust Variant. Hans David cites Schenker´s repeat marks and invalidates them (24).
2.2.6. Hans David, 1926
Hans David, in his seminal essay on the Chromatic Fantasia and Fugue from 1926 (25), writes of the Rust Variant:
Wir haben hier offenbar eine frühere, eine erste Fassung der Fantasie vor uns. Der Sinn der späteren Umarbeitung ist unschwer zu erkennen. Bei der Besprechung des Werkes zeigte sich in hohem Mass ein Fortschreiten des Geschehens, eine musikalishce Entwicklung. In den späteren Teilen der Komposition stand solche Zielstrebigkeit offenbar bereits zur Zeit der ersten Niederschrift fest. Der erste Teil hingegen war hier in die allgemeine Bewegung noch nicht einbezogen; die stetige Achtelbewegung etwa—um einen markanten Zug herauszugreifen—machte die Empfindung einer Stufenfolge, eines Fortschrittes geradezu unmöglich. Dieser Teil hatte nicht den Grad von Singularität, den die späteren Abschnitte bereits besassen: es blieb ein Missverhältnis. Bach tilgte es, indem er die Stücke des Teiles als verschiedenartige Abschnitte, denen jeder eine höhere Stufe, ein späteres Stadium als der vorangegangene erreicht, fühlen liess, auseinanderlegte. Jetzt erst wurde die Fantasie ein homogenes Gebilde. (26)
“We obviously have an earlier, first version of the Fantasia before us here. It is easy to see the reasons for the later reworking. In the discussion of the latter we have noted a steady musical development. In the later parts of the composition this tendency was already true of the first version. But the first continuous eight-note movement—to mention such a noticeable feature—makes this kind of steady forward progression impossible. This section did not have the unique characteristics of the rest of the work: it remained a disproportional relationship. Bach remedied this by presenting the sections as separate units and by making sure that each section became a further stage of development than its preceding one. Only then did the Fantasia become a homogeneous entity.” (Translation by Rosalyn Tureck)
In this study of the Chromatic Fantasia and Fugue, David, repeats the opinion, initiated by Forkel in 1802, of Bach laboring step by step (27). For Forkel, in his biography of Johann Sebastian, writes:
“We have several times spoken of the great care with which Bach endeavored all his life long to improve his works. I have had opportunities of comparing together many copies of his principal works, written in different years, and I confess that I have often felt both surprise and delight at the means which he employed to make, little by little, the faulty good, the good better, and the better perfect…” (28)
However, twenty-one years later, in 1945, the footnote in the Bach Reader to Forkel’s assertion reads as follows:
“Forkel happened to know a number of Bach’s compositions in various readings. Taking all of them for authentic (although some of them were not), and generalizing from what he had seen, he arrived at an impression that was not quite correct. The majority of the works by Bach remained unchanged after their original composition, and even such changes as Bach did make were usually of minor extent and importance.” (29)
This later opinion is likely to have been based on the famous article published in 1938, by Richard S. Hill (30), in which Hill points out that Forkel erred in his chronological assumptions, as well as in several of the evaluations of the manuscripts that he had acquired. Hill writes:
“Cahier 2 (Hoffmeister and Kühnel Collected Keyboard and Organ works, S. Bach’s sammtl. Klavier- und Orgelwerken) follows Forkel’s advice and makes use of what Forkel thought were late revised editions of Bach’s works—in reality these were the earliest manuscripts… The errors [in the publisher’s announcement for Cahier 2 (31)]… are all so typical of Forkel that there can be no doubt that the information—as well as the faulty manuscripts—came from him…” (32)
Hans David recognizes, as does everyone, the disproportion between mm. 3–24 and the rest of the Fantasia. He perceives that it “did not have the unique characteristics of the rest of the work” (33). He goes so far as to say that the continuous eight-note progression “makes it impossible to relate to the rest of the Fantasia”.
Yet these perceptions did not provoke the deeper question—does this incongruous, simplistic section, tacked on to the rich texture and imaginative structuring of the main body of the Fantasia, really emanate from the same brain? David perpetuates the view in his imaginative compounding of Forkel’s thoughts—derived from a mistaken chronology—about Bach’s style of creativity. Hans David writes: “Bach remedied this by building the varied sections by making sure that each section became a further step of development. Only then did the Fantasia become a homogeneous entity.” (34)
But creative minds do not work in this plodding, linear style. Although Bach was not a total stranger to revision, he was not the type to rework “to make sure” (35) that each section became a further stage of development than its preceding one. This style of step-by-step labor belongs in the schoolroom.
Hans David condemns Schenker’s repeat marks: “Schenker places repeat indications for mm. 3–24; he has not support for this in any manuscript.” (36) David is correct; he goes on to say: “…such deviations must be strongly condemned…” However, in his effort to perceive some reason for Schenker’s imposition upon the Variant, he hits upon the hope that it is due to its weakness: “Certainly the development is so weak that a repeat might work; … even if the repeat were musically justified, one would have to ask for distinct proof of the origin of the idea.”
A query comes to mind here: Does a weak section become stronger by repetition? Such an assumption is, in itself born of weakness. In actual fact, a repetition of a weak section simply emphasizes its weakness.
The plate number of the Chromatic Fantasia and Fugue, as quoted in David´s article is incorrect. “.. Peters Plate number 1512. The greatest section of the Fugue was made from the old plates, PN 72, printed on pp. 9–12.” (37) PN 72 is the plate number for the fifth Partita in G-major, BWV 829, Cahier Six in the Hoffmeister & Kühnel Bach Series (38). The original number of the old plates is PN 74; these do not make up the “greatest section of the Fugue” (39). They represent four pages, exactly the first half of the movement. The succeeding four pages of the Fugue are made from the new plates, PN 1512, necessitated by Griepenkerl’s changes in the text (40).
2.2.7. Wolfgang Schmieder, 1950 (41)
903a Fantasie d-moll
(Variante zu Nr. 903)
BGA XXXVI, 219 – EZ Kothen etwa 1720.
Vermutlich die Urform der Fassung No. 903
(Presumably the original form of the setting No. 903)
Abschrift. 11 Seiten Hochformat (“J. L. A. Rust, Bernburg, 1757”) aus d. Nachlass Wilh. Rusts. Letzter bekannter Besitzer: Karl v. Vietinghoff, Berlin.
(From the estate of Wilhelm Rust. Last known owner: Karl v. Vietinghoff, Berlin)
Schmieder is somewhat more cautious than his predecessors: he begins his chronological assessment of the Rust Variant with the word “presumably”.
The major interest in Schmieder here is his identification of the provenance of BWV 903a. He also pinpoints the last known provenance to Karl v. Vietinghoff. I traced Vietinghoff and learned that he was deceased (ca. 1936), but that a daughter was reported to be alive. I followed the trail but it led nowhere: she was not to be found.  Schmieder does not attempt a scribal identity nor does he mention the lost cover.
2.2.8. Hermann Keller, 1962 (42)
Die wohl älteste Gestalt der Fantasie findet sich in einer Handschrift F. W. Rusts, die sich dieser im Jahr 1757, also im Alter von 18 Jahren angefertigt hat. Sie ist in unserer Ausgabe im Anhang I ganz mitgeteilt. Der Anfang ist von der späteren, hier im Hauptext mitgeteilten Fassung völlig abweichend, von T. 21 ab aber nur in Einzelheiten anders. (43)
“What is likely to be in the oldest form of the Fantasia can be found in a manuscript of F. W. Rust produced by him in the year 1757, that is, when he was 18 years old. It is reproduced in full in our edition, Appendix I. The beginning diverges completely from the later version, reproduced here in the main text, but from bar 21 it diverges only in details.” (Translation by Rosalyn Tureck)
Keller offers no support for his ascription. He, also, assumes that the Rust version is “likely to be the oldest form” and that BWV 903 is the “later version”. At no time has any clear evidence been produced that the Rust Variant is an early version. The notion, rightly or wrongly initiated by Griepenkerl, has become fixed by the repetitions of posterity into an unquestioning acceptance, emerging here with Keller, well over a century later, into a concrete attribution.
No report based on the investigation of the Rust manuscript has surfaced from the period when the manuscript was available for a proper investigation. Moreover, it is not clear, according to the editors previous to Naumann who, if anyone, did actually see the original cover containing the name of J. L. A. Rust. It is not even clear who saw the music text of the Rust manuscript. One deduction seems appropriate—Griepenkerl, who was the first to include the Variant in his edition, would surely have seen the score, in preparation for his edition of 1843/44. It is to him that we owe the opinion that this Variant is the “oldest form” of the opening two pages of the Chromatic Fantasia (44). That this opinion was repeated by following editors does not establish unequivocal verification founded on the document itself and on an investigation of the circumstances of its appearance. Numerous instances of repetition throughout history based on assumptions, or even mistaken information originally put forward, have occurred in other branches of study and it may well be that this is such an instance.
Considering all the facets of the case, the ascription of F. W. Rust as scribe for BWV 903a remains in the area of subjective assumption.
2.2.9. Georg von Dadelsen and Klaus Rönnau, 1970 (45)
Eine andere Fassung der Fantasie (BWV 903a) bot die verschollene Hs. aus dem Besitz von F. W. Rust, die “Bernburg 1757” datiert war. Ihre Abweichungen betreffen im wesentlichen die 18 Takte 3–20, wofür sie 21 ganzlich andere Takte bringt. Wir teilen diesen Schnitt nach BGA 36 als Variante im Anschluss an die Fuge mit. Die betreffenden Takte sind gradliniger, weniger abwechlungsreich als der entsprechende Abschnitt der Hauptfassung, aber dennoch von grösser Wirkung. Da man sich die bekannte Fassung sehr wohl als Weiterbildung dieser Variante vorstellen kann, das Umgekehrte jedoch kaum möglich ist, gilt diese Fassung als die früheste Form des Werkes. Allerdings stützt sich diese Ansicht nur auf diese Partie. Die übrigen Sonderlesarten der Rustschen Hs sind dagegen in keinem Fall als Frühfassungen zu interpretieren. (46)
“Another version of the Fantasia (BWV 903a) existed in the lost manuscript dated ‘Bernburg, 1757’ belonging to F. W. Rust. Its variants concern mainly bars 3–20, instead of which it gives 21 totally different bars. We give this Variant from B.G.A., 36 in our appendix to the Fugue. The bars in question are more direct and less complex than the corresponding section in the main version, but nevertheless of great effect. Since it is easy to imagine the well-known version as a further development of this Variant, and the opposite is hardly possible, this Variant must be the earliest version of the work.” (Translation by Rosalyn Tureck)
Georg von Dadelsen and Klaus Rönnau are also bound to the traditional notion that the Variant is the “earliest” version written by Bach. Their deduction that the simper version can hardly have followed BWV 903 is arrived at through the same linear process of thought that produced the assumptions of the nineteenth century editors. But they are alone in their assessment of the Rust Variant as being “von grösser Wirkung”—greatly effective. 
Two other recent editions of the work hold similar views of the Variant 903a, which remains being considered as an early version of the BWV 903. Firstly, the Wiener Urtext Edition, edited and commented by Ulrich Leisinger, with fingering and notes on interpretation by Michael Behringer, notes that “there are at least three simpler versions whose authenticity cannot easily be doubted. They may therefore be understood as early stages of the composition” (Leisinger, ed., ‘Preface’, IX). The Rust Variant, identified here as the “earliest” of such stages is provided as the first appendix to the main version—although only the first 31 bars thereof, on the basis that the differences between them are “comparatively small”, and that these “have been documented in a special section of the Critical Notes thus making it possible for the performer to reconstruct completely the early version” (Leisinger, ed., ‘Preface’, X). The other two earlier stages are the versions contained in the manuscript copied by Johann Tobias Krebs and a “collective manuscript compiled by Forkel (source F)” (Leisinger, ed., ‘Preface’, IX).
Secondly, the edition prepared by Uwe Wolf for Bärenreiter states that “the early version of the Fantasia (BWV 903a) is accesible today in a single manuscript prepared in the 1730s for the Darmstadt Court Chapel. Another copy of this version was extant in the nineteenth century and served as the basis of both the Peters edition and that of the old Bach Gesamtausgabe. The early version has a radically different opening section and an unusually large compass, from contra A to d3” (Wolf, ed., ‘Preface’, IV). It also mentions an “intermediate version closely resembling the main version” which “survives in a copy prepared by Johann Tobias Krebs the Elder”, and “in a manuscript copy written in the hand of Samuel Gottlieb Heder” (Wolf, ed., ‘Preface’, IV-V). Editor Uwe Wolf recognizes, in a fashion similar to that of the Wiener Urtext Edition's editors, that “performers will likewise find it interesting to have this famous work available in three self-contained evolutionary layers” (Wolf, ed., ‘Preface’, V).