Homepage of WTC II MSS
London Autograph
BWV 891,2 (b.67)

Manuscript Sources of J. S. Bach's Well-Tempered Clavier II

Yo Tomita

Group A Manuscripts

Tradition A is a counterpart of Tradition B: it attests totally different histories of compilation and revision from those of L. As illustrated graphically in Critical Commentary, Vol.1 (1993), page 6, its basis was S, the lost autograph set, which not only served as an exemplar for L, but also for many copies of its own Tradition. This manuscript must have contained most fascinating information about Bach's process of revisions on both early discarded readings and the subsequent reworkings that he entered in S, as opposed to L. While the great majority of such information is no longer retrievable, some of it can be considered as being reflected in the subsequent copies. Among these is P 430. This MS can be considered as the most important non-autographic source for WTC II, because it is the earliest surviving copy stemming from S, and many other manuscripts were both copied from it and corrected retrospectively with reference to it. The sources originating from P 430 form Group A1.

P 430 was made by Bach's future son-in-law, J. C. Altnickol (1719-1759). Two lines of added inscriptions, in different ink and calligraphy, are written at the end of the volume: 'Scr: Altnickol.' and 'ao. 1744'. This informs us of the scribe and the date that is associated with the MS. 1744 was the year in which Altnickol matriculated at Leipzig University and became Bach's pupil. This biographical fact suggests that the manuscript was one of the first tasks assigned to him, as Bach's pupil, and according to Dürr, the characteristics of his handwriting in P 430 show the earliest known record of his calligraphy.

Altnickol's other complete copy of WTC II, P 402, also survives. This manuscript, dated 1755, is not copied from P 430, but as it shares a significant number of the same unique errors as P 430, so that Altnickol must have used the lost intermediate source A' to write both MSS. It is probably the case that P 430 was left in Bach's household, and Altnickol kept A' instead, when he left Leipzig in 1748.

A closer inspection of P 430 reveals that Altnickol's task was not to reproduce a copy in a straightforward manner, however. Doubtless P 430 is an extremely carefully prepared copy, but on the other hand, certain movements show a trace of Altnickol's apparent difficulties with the text, manifested in the form of an excessive number of errors of restricted kinds. The analysis of these suggests a probable cause: a situation where Altnickol seems to have been instructed by Bach to transform certain aspects of the pieces at copying stage, such as the change of R.H. Clef from the treble clef to soprano (Pr.C and Ab) and the metre and the associated note-value of movements (Fg.bb and Pr.b).

The other interesting features of P 430 include traces of 'thorough usage', seen in later amendments in various different hands and inks. Under normal circumstances, the ascertainment of authorship is extremely difficult, due to the lack of information gleaned from the study of calligraphy. Thus we should also take into account the musical and notational significance of each amendment, and compare the result with the chronology of readings witnessed in the genealogy of sources. One of the relatively secure criteria for our calligraphic study is the analysis of crotchet rests: here Altnickol's initial shapes can be easily distinguished from those by Bach and by Friedrich August Grasnick (ca.1800 - 1877). It is worth noting that the shape of Altnickol's crotchet rests gradually, with the passage of time, becomes closer to that of his teacher, as is shown in his subsequent copy, P 402. Thus we should be careful when attempting to determine the authorship of later entries close to 1748.

The calligraphy of sharps, '#', also contains some information about the scribe. Interestingly, Altnikol has two distinctive forms of sharps. The former is the standard one, which he basically retains in P 402. This neat symbol is used for keysignatures and the accidentals where he could have written with confidence and intention. The latter rough type is the result of writing at greater speed and with haste. There are also cases where the shape of accidentals is severely affected by the lack of available space to write them comfortably. This is commonly caused by his lack of insight at the copying stage in reserving room for them, or by his later deciding to add them. Occasionally we can identify the sharps in Bach's hand as well.

The later amendments in P 430 can be classified according to their nature into the following categories: (1) corrections at either copying or proof-reading state; (2) introduction of deliberate variant readings; (3) incorporation or updating of the variant readings from the other Tradition. Each category of activity may cover several years and involve various people, and furthermore, such revisions most likely overlap each other so that the accurate assessment of their chronology is extremely difficult. Yet it is worth noting that many of the unique, inspired readings are entered by Bach into P 430, which source defines the important character of A1.

Let us place P 430 into the genealogical context. Our text-critical study suggests the following tree:
Diagram of Sources, A

The only way left for us to acquire information about the lost S is to study in most minute detail the surviving sources which derived from it. For this purpose, it is considered sufficient to examine five principal sources, namely P 430 and P 402, RCM 743, Poel.33,2 and PM 5697. The first two stemmed from Altnickol's personal exemplar, A' (lost), and the remaining three from S, independently, via their own lost intermediate sources. All the rest of the sources in A derived from P 430. I shall discuss briefly each source in the context of its text-critical value, and then return to discuss the mainstream, A1, as it seems to be the correct chronological sequence from the point of textual maturity.

The interpretation of Altnickol's second copy, P 402, is still controversial. There is no doubt that at least for the majority of movements, the same Vorlage of P 430 must have been used in its preparation. Yet even allowing for the inclusion of Altnickol's errors and Bach's revisions in P 430, we occasionally encounter the following problems when attempting to place P 402 into a single context, namely that it contains:
  1. the oldest known reading (e.g., Pr.g#, bb.44-45)
  2. the readings of L instead of the characteristic readings of A (e.g., Pr.d, final cadence; Pr.F, b.28; many more cases also shared with PM 5697)
  3. the same reading as non-Altnickol sources of A, which is a more advanced state of reading (e.g., Pr.g#, b.20: A,3 [except RCM 743])
  4. the most advanced state of reading (e.g., ornaments of Pr.a, bb.16 & 32; and figuration of Fg.E, b.12 and Pr.A, b.17)
To clear up all of the above problems, there are certain conditions to be met: first of all, A' indeed contained the earlyiest reading, but Bach must have given specific instructions to Altnickol (when he made P 430) to make alterations (Pr.b is an excellent example); secondly, it is possible that a tiny proportion of early readings in A' was updated between 1744 and 1748; thirdly, A' was most likely merely one of several copies made from S, thus the change made to one source might or might not have been taken into the other contemporary copies; and finally, some of the new, inspired readings may be Altnickol's own. The respective likelihood of each of the above proposed conditions are yet to be measured, and merit future study.

The recently discovered source, RCM 743, is an English source. Inside its front cover is the inscription "J. W. Windsor | Novr 30 | 1801 | Bath". It is classified with P 402 as Group A2, as it shares with the latter source unique characteristics, in that Fg.bb follows the metric convention of L/A1, while Pr.b follows that of A3. Detailed examination shows that its text is rather closer to the common Vorlage of A1 and Poel.33,2. Although the MS is heavily changed notationally, there is insufficient evidence for interpreting that the text is contaminated as a result of eclectic editorial work. Far more significant is the fact that it seems to contain both unique variant readings of what appear to be the authentic earliest readings in S, such as Fg.d#, bb.17-19, and ones shared with Poel.33,2. There are also many other unauthentic readings, which may also be explained by the likely situation of the scribe (or for this matter, any scribes of the same ancestor sources) offering conjectural renderings of heavily corrupted or revised text.

The remaining non-Altnickol sources, Poel.33,2 and PM 5697, are classified as Group A3, which is characterised by half note-value notation of Fg.bb and Pr.b. There are a number of shared unique readings (mostly errors) in both sources, suggesting their close relation. While the text of PM 5697 appears to have been slightly contaminated with that of P 402 (or its common source A' ?) and apparently edited, Poel.33,2 seems to attest one particular state of S. This unique quality is extended to the way movement headers are written, which seems to reflect Bach's policies in distinguishing a group of sources of similar textual maturity, or even two different autograph set, L and S. To what extent we can rely on its text as authentic is questionable, as the text in the hand of J. C. G. Bach is full of errors -- mainly omissions of symbols, such as accidentals, ties and ornaments, and pitch errors. Nonetheless, it can also be said that this is a proof of genuine, uncontaminated text. Based on such expectations, it seems significant to find the following:
  1. two readings from different traditions being superimposed (e.g., final cadence of Pr.d and Fg.f#; Fg.Bb, bb.5-6);
  2. unique authentic reading of this group;
  3. unique unauthentic reading possibly caused by unclear text, as a result of Bach's revision in S (such as Pr.b, b.8).
Another noteworthy point about the text is that among numerous errors in Poel.33,2, there are not a few readings which are also recorded in P 430 as a.corr. readings. From the statistical point of view, this is significant. It is safe to assume that the errors originated in their common Vorlage, most likely S itself. As these errors usually do not occur in P 402, we can assume that the same corrections were also entered in A'.

We return to the mainstream, A1. I have allowed the presence of a lost intermediate source, A'', to the majority of copies derived from P 430. This is because there are many isolated errors of various kinds in P 430 which seem to have been corrected elsewhere, and this state of MS became the basis for A1. The natures of such errors in P 430 suggest two likely sources of origin -- inherited errors from the Vorlage and Altnickol's own errors when writing P 430. There is further evidence of P 430 containing numerous subsequent textual alterations, which range from corrections to the introduction of unique, often inspired, variant readings -- some of which are clearly identifiable as being in Bach's calligraphy. This is logical, too, as Altnickol studied for four years under Bach, presumably using this MS. The majority of these changes is also reflected in A1.

Within A1, there is the unresolved issue of the chronology of textual alteration, as some portions of A1 give a.corr. readings, others giving p.corr. readings. As will be covered in the discussion of individual movements, I find no convenient way at the moment to classify the chronology of readings from the study of surviving sources as a whole, as different pictures of chronology are portrayed in different movements. The solution may be available when we are able to pursue a scientific measurement of the likelihood of textual contamination and to confirm the existence of further contemporary sources, on which the sporadic transfer of revised readings took place, in Bach's household.

The revision history of P 430 continues even into the 19th century. The only identifiable hand is that of Grasnick, who also wrote several WTC II manuscripts, P 546 and P 549. It appears that he compared P 430 with Am.B.57, edited P 430, and subsequently wrote P1146 in which he listed the noteworthy variant readings he found in Am.B.57. It must have been him who wrote 'X' marks in both P 430 and Am.B.57. Despite the fact that in P 1146 he gives a list of variants found in twenty movements, the X's in Am.B.57 are restricted to eight movements only, viz. Fg.C#, Pr.c#, PrFg.D, Pr.d, Pr.Eb, Pr.d# and Pr.Ab. On the other hand, in P 430 Pr.C# is the only marked movement -- the one he copied in full (with some mistakes) in P 1146. There is also a note, in his hand, pasted to the first page of P 430, entitled "2ter Theil v Seb Bachs wohltemp. Clav verglichen mit Kirnbergers Handexemplar" in which he summarises the comparison of the two sources.

It seems that his basic editorial strategy was to supplement the symbols present in Am.B.57 which were missing in P 430. These are mainly rests, ornaments and accidentals. Occasionally he wrote annotations only, as in Pr.F#, b.80. Among the transferred readings are some strange errors in the manuscript circle of Am.B.57. Whatever editorial procedure he chose, his preference of readings seems to have been determined by his limited musical understanding and taste. With all the evidence gathered, including that of his general attitude towards the manner of amendments, it is reasonable to infer that Am.B.57 belonged to his teacher, G. H. Moering, the pupil of Kirnberger, and that Grasnick himself owned P 430.

The amendments by Grasnick were probably the last, and can be dated approximately 1820. Thus it must be the case that all of the amendments made by his hand were unavailable when P 204 was initially copied from it, in 1781. There are numerous instances of what appears to be a transfer of readings from those added into P 430, by Grasnick, to P 204, as further additions. This is unclear if the reference is made directly.

Top Contents Homepage

Last modified: 15 September 1996