Authored by: Yo Tomita

The Routledge Research Companion to Johann Sebastian Bach

Print publication date:  November  2016
Online publication date:  November  2016

Print ISBN: 9781409417903
eBook ISBN: 9781315452814
Adobe ISBN: 9781315452807




Any consideration of Bach manuscripts must begin with the work of two scholars: Georg von Dadelsen (1918–2007), who broke new ground in Bach scholarship by dating Bach’s manuscripts through analysis of the handwriting of Bach and his copyists; 2 and Alfred Dürr (1918–2011), who was able to establish a new chronology of Bach’s vocal works composed in Leipzig by studying watermarks and identifying the individual copyists. 3 For those who accepted the validity of the old chronology proposed by Philipp Spitta (1841–94) in the late nineteenth century, which until then had been considered unassailable, it was an earth-shattering realization that the foundation of their scholarship had been removed. A half-century has passed, and although a number of refinements have been made to the findings of Dadelsen and Dürr (and others), their theories and methodologies have remained broadly unchallenged. There is, however, every reason to believe that future generations will develop new techniques and methodologies that will re-evaluate existing evidence and discover new perspectives in Bach research.

 Add to shortlist  Cite


Any consideration of Bach manuscripts must begin with the work of two scholars: Georg von Dadelsen (1918–2007), who broke new ground in Bach scholarship by dating Bach’s manuscripts through analysis of the handwriting of Bach and his copyists; 2 and Alfred Dürr (1918–2011), who was able to establish a new chronology of Bach’s vocal works composed in Leipzig by studying watermarks and identifying the individual copyists. 3 For those who accepted the validity of the old chronology proposed by Philipp Spitta (1841–94) in the late nineteenth century, which until then had been considered unassailable, it was an earth-shattering realization that the foundation of their scholarship had been removed. A half-century has passed, and although a number of refinements have been made to the findings of Dadelsen and Dürr (and others), their theories and methodologies have remained broadly unchallenged. There is, however, every reason to believe that future generations will develop new techniques and methodologies that will re-evaluate existing evidence and discover new perspectives in Bach research.

The aim of this chapter is three-fold: first, to explain systematically the multiple disciplines that have to be used in the study of manuscripts; second, to review the evolution and development of methodologies used by the scholars who have shaped the present form of scholarship, and to chart outstanding problems that have yet to be resolved; and third, to offer some ideas of what future research might entail and in what way scholarship might develop. Because numerous and disparate methodologies are used in the study of Bach manuscripts, the discussions that follow will take nothing for granted, but will describe and define each one as it relates to understanding and reproducing, with as much accuracy as possible, Bach’s intentions in the manuscripts that contain his music.

Manuscript Study

Bach research has traditionally placed a great emphasis on studying manuscripts. The fact that most of Bach’s works have been transmitted in manuscript form is the obvious justification for such studies. But, crucially, there are broader musicological research goals than simply deciphering the music of a manuscript. Not only is there the music but also evidence of both writer(s) and user(s) that witness to the origin and dissemination of the composition(s). A manuscript may contain additional important details, such as interventions by later owners, individual and institutional, auctioneers and antiquarian music dealers. Whether or not such information can be retrieved and decoded successfully depends on the methodology and research techniques used. Even if the significance is not immediately apparent the information should nevertheless be recorded, opening the possibility for future discovery. Scholarly advancement moves forward in this perpetual cycle, building on the discoveries of one’s predecessors.

Every music manuscript contains a wide range of information manifested not only in its written contents, such as the various inscriptions including the title, movement headers, or the composer’s name and, of course, the actual musical contents in notational form, but also in its existence as a whole, which reveals the circumstances of its production. This includes the physical makeup of the score and/or a set of parts, the identity of the writer, when and why it was produced, and evidence of its subsequent ownership and use. If the manuscript was written by the composer himself, considerable insights into his creative process might be gained through a careful exploration for traces of composition and revision that it might contain. Manuscripts are indeed a treasure-house of information, and from them one may learn many fascinating historical details. A manuscript may offer clues that disclose the date of composition and performance(s), or information about its background, the person(s) who inherited it, and how it has been used or appreciated by subsequent musicians and collectors. This kind of information may not be accessible from the study of the surviving letters and documents alone. In this way, manuscripts may provide fresh perspectives on Bach’s biography as well as the reception history. Manuscript studies will therefore continue to attract scholars eager to unveil quiescent pieces of evidence waiting to be brought to light.

Approaches and Terminologies

The study of a manuscript usually begins with an assessment of the manuscript itself as a physical artifact. This involves examining diplomatic evidence, that is, all its material features such as paper; the implements used in its production such as quill, pencil, rastrum, ruler and knife; the types of ink used; foliation of paper and binding; as well as any significant features in the handwriting specific to time and region, each of which supplies an independent set of chronological and locational information with which to profile the circumstances of the source’s production. Although this phase of examination, when conducted on a single source, may not immediately produce definitive evidence about its origins, it is vital that one takes this step properly from the outset because the evidence garnered through this process becomes the basis for interpreting evidence acquired through subsequent phases of examination. Quite separately, this step needs to be undertaken as part of a large schematic research examining all related sources. The greater the volume of diplomatic evidence gathered, the greater the possibility of ascertaining the origins and contexts in which the sources were made.

The next phase of inquiry is text criticism, the evaluation of the written contents using philological methodologies. The procedure normally begins with an assessment of the broader issues, such as obvious notational features, and gradually narrows down to a detailed discussion of the specifics of musical content, such as corrections and revisions, incrementally building on observations made in the previous phases of study. When examining a work that is transmitted in multiple sources, one of the primary aims is to establish which source gives the closest readings to the original even if the original is no longer extant and has to be reconstructed hypothetically. Ultimately, the study should establish a firmer ground to assess the content of a manuscript and establish what the composer intended when the document was being written.

It is not always straightforward to retrieve the information one expects to find. To successfully locate the information and make sense of it, it is usually necessary to examine widely and systematically not only the source in question, but also all the other sources that are considered related for specific enquiries. Furthermore, the sources should be evaluated both externally (the physical makeup such as paper, ink, foliations of the manuscript, etc.) and internally (the manner in which they were written as well as the written contents of the manuscripts). Imagine that each manuscript represents a collection of various tiny fragments of knowledge: it resembles a multi-dimensional jigsaw puzzle—each fragment may provide crucial information for reconstructing the image, but the same piece of information may also support the reconstruction of another picture. For example, a variant reading in one source can provide evidence for tracing the transmission route of related manuscripts, but the same information can also be seen as evidence for reconstructing the profile of the scribes (how reliably they reproduced their model, which aspects of the musical text they altered and why). Information retrieval from the music manuscript needs to be conducted in such a way that allows for subsequent analyses and verification.

Locating Sources

The usual starting point for Bach scholars has been to consult the relevant volume of the NBA Kritische Berichte that attempts to trace, describe and assess all the known manuscript sources for specific works of Bach. 4 These volumes provide the background to the decisions behind the edition of music in the specific volume of the NBA. However, one must be aware that, because the NBA project ran for over half a century and involved many scholars, the quality and the state of scholarship in each volume are not uniform, which has resulted in many of the earlier volumes requiring major updates. More recently, scholars have increasingly been using online databases of sources, such as the Source Catalogues of Bach-Digital (2008– ), 5 which succeeds the Göttinger Bach-Katalog (1999–2008). 6 In addition to the basic search facility, such databases allow filtering the information by scribe, approximate date of copying, Bach’s compositions contained in the source, and other basic source information. The databases also offer high-resolution color scans of Bach’s original manuscripts held in public libraries in Germany. Although the amount of detail available in online databases is limited—and for this reason the NBA Kritische Berichte retain their usefulness—the value of the databases is unquestionable, as their data can be frequently updated, errors corrected and new records added. They have become the preferred initial source reference tool for most Bach scholars. In the past, one had to write to the holding library to request microfilms or photocopies to obtain reproductions of sources that were not commercially available in facsimile 7 or microfilm/microfiche. 8 The Bach-Archiv Leipzig library possesses a comprehensive collection of photographic reproductions of manuscripts. Today, color scans of manuscripts can be obtained on CD-ROM or downloaded from the internet, and, once scanned in this manner, they are often made available to other readers—sometimes free of charge—from the library’s website. 9 Because of the speed at which digital images are becoming available on various websites, it is increasingly difficult to keep track of the newly digitized sources. 10

To locate sources that have not yet been reported in Bach literature, other channels such as RISM and individual library catalogues that may have escaped the notice of Bach scholars need to be investigated. It is, of course, possible to stumble upon new sources in archives or auction houses purely by accident. Vague descriptions are common in nineteenth-century catalogues of sources and auctions, and it requires considerable knowledge about the many aspects of music manuscripts to develop an instinct for evaluating what the cataloguer was attempting to describe, and piece together numerous minute fragments to reveal the hitherto invisible historical information.

Classifying Sources

Sources can be classified in a number of ways, determined by the nature and purpose of the enquiries. Classification of sources can be confusing; any study of a particular composer tends to generate its own distinct source types and tailor its jargon accordingly, which has resulted in an extensive accumulation of terminology. 11 Many of the source type descriptions—sketch, draft, composing score, fair copy, etc.—were devised for the purposes of scholarly discourse, and were not necessarily used by the composers themselves. It is therefore of utmost importance that in source scholarship a distinction be made between our later categories and those of Bach’s day.

Primary and Secondary Sources

The common practice in manuscript studies is to divide sources into two categories: primary and secondary. Primary sources are those shortlisted as priority subjects because they are believed to contain the original information, so that the remaining sources that do not contain any original information from the time period of enquiry, that is, the secondary sources, can be set aside from the main target of research.

The majority of studies of music manuscripts are conducted with a view to producing reliable editions of music. One of the most important tasks of the editor is to establish the composer’s intended text through the examination of the surviving sources. 12 The primary sources consulted for this purpose include everything that directly reflects the composer’s intentions. This includes scores written in his own hand as well as performance parts supervised by the composer. These are more commonly referred to as “original manuscripts” (see the following section). Also included in this category are prints published during Bach’s lifetime, especially by Bach himself (i.e., “original prints”), that may contain information not found in the surviving manuscripts. 13 In reception studies, the category of primary sources is accordingly broadened to include sources of the period on which the study is focused. For example, a manuscript copy of Bach’s Schübler chorales (BWV 645–650), once owned by Mendelssohn (GB-Ob, MS. Deneke Mendelssohn c.71), would be regarded as a primary source by someone studying Mendelssohn’s involvement with the publication of the work, whereas someone preparing an edition of the Schübler chorales would consider it a secondary source. 14

The classification of sources is an important part of a scholar’s work, and false assessment at this stage could jeopardize the outcome of the discussion in a catastrophic way. A well-known case is Friedrich Smend’s evaluation of two specific manuscript copies of the B-minor Mass, Source C (D-B, Am. B. 3) and Source D (a three-volume manuscript consisting of D-B, Mus. ms. Bach P 572, P 23 and P 14). Smend considered both these manuscripts to be primary sources dating from Bach’s lifetime (i.e., written before the autograph was modified), whereas later scholars found that they were copied after Bach’s death (the alteration of the autograph having already begun). The damage, however, had been done and, consequently, the NBA edition of the Mass incorporated numerous posthumous additions as if they were Bach’s own. 15

Autographs, Copies and Original Manuscripts.

When the writer is the author of the work, the manuscript is termed autograph (or less often holograph). 16 All the other manuscripts are thus called copies. Copies (Abschriften) usually contain a greater number of errors of various kinds. Actually, it is a somewhat shocking revelation to find that a number of the original parts of Bach’s cantatas contain uncorrected errors, some of which so serious, that they surely must have caused problems when they were used for performance. 17

In Bach literature, there are a few other terms that are commonly used to distinguish the finer nuances between sources. Partial autograph (Teilautograph) is a hybrid type of which there are two possible materializations. The first is a manuscript that was written by more than one writer, one of which is Bach himself. A typical example is the score of Cantata 61 (D-B, Mus. ms. Bach P 45, Faszikel 6), which was started as a fair copy by Bach’s copyist, with Bach himself taking over from the second movement. 18 There are also cases where Bach embarked on making the fair copy, which he then passed on to his assistant for completion. The score of the St. John Passion (D-B, Mus. ms. Bach P 28) is an interesting example. Having copied the work neatly up to page 20, in mid-March 1739 Bach suddenly stopped copying, possibly because he learnt that the Passion performance had been canceled for that year. 19 The manuscript was completed in 1749, by one of Bach’s principal copyists, Johann Nathanael Bammler (1722–84), who seems to have used the draft score of 1724 that contained revisions made before 1739. 20 The second manifestation of a partial autograph is a copy that contains Bach’s later additions or corrections. Examples are both numerous and wide-ranging, and include the original score—a revision copy—of BWV 875/1 and a sales copy of BWV 1021 (see discussion later in this chapter).

Some confusion exists over Bach’s copies of other composers’ works, either the copies he made himself or those he owned, but did not copy: they should never be referred to as Bach’s autographs, even if they are entirely in Bach’s handwriting. When discussing Bach’s transcriptions of other composers’ works, the distinction becomes even more confusing, as a transcription may denote anything from a straightforward re-setting for a different instrument or instruments to more involved re-compositions of a piece. In principle, a straightforward re-setting for a different instrument should be called “copy” or “transcription manuscript.” This applies when discussing the non-extant manuscript in Bach’s hand of the Aria (BWV 587), a transcription of Couperin’s Trio movement from the Suite L’impériale for two violins and continuo. On the contrary, when discussing the score of the D-minor Concerto for organ (BWV 596; D-B Mus. ms. Bach P 330), which is a more involved re-composition of Antonio Vivaldi’s D-minor Concerto for two violins and cello from L’Estro Armonico, Op. 3, No. 11, RV 565, it should be referred to as “autograph” (of Bach’s arrangement), and not “copy” (of Vivaldi’s composition). 21

The original manuscript (also original score and original parts) is another conceptual term used in Bach scholarship. It refers to a manuscript prepared for Bach’s own use, his personal copy for that matter, which could have been made either by Bach himself or another copyist (the question of penmanship is immaterial here). However, as Bach himself did not make any remarks that would indicate that the manuscript was his personal copy, this can only be found out through studying the surviving sources systematically and carefully. One of the best-known examples of this kind is the so-called “London autograph” of WTC II (GB-Lbl, Add. MS 35021), which is a partial autograph, mostly in Bach’s hand, but some movements are in the hand of his second wife, Anna Magdalena Bach, who assisted her husband in the production of fair copies. 22 A question such as that concerning the division of their labor becomes an important issue when evaluating the source: Anna copied from Bach’s revision copy that needed to be produced in fair copy, while Bach took on the portion he felt he still needed to revisit from the compositional aspect. The concept of original manuscript is crucial when discussing the sources of Bach’s ensemble music, particularly those for a vocal ensemble with an orchestra, such as the sacred cantatas. Bach composed his music as scores, from which sets of parts were then produced, often by his copyists. When he subsequently inspected the parts, he supplied performance-related indications, such as slurs and ornaments, but also further developed some of the musical ideas. And, normally, these were not transferred to his own scores. For this reason, in Bach scholarship it is more productive to discuss the sources from Bach’s perspective rather than from our own arbitrary distinction between autographs and copies.

Compositional Stages and Neatness of Handwriting

Manuscripts are customarily described in pre-classified categories of notional compositional stages from general appearance, namely, sketch (Skizze), draft (Entwurf), composing score (Konzeptschrift) and fair copy (Reinschrift). 23 In Bach’s case, the first two of these types do physically exist, but only within composing scores, as can be observed in the composing scores of some cantatas, for which Bach did not produce fair copies. It appears that his standard practice was to write out a score directly, without relying on sketches or drafts, 24 by seeking to develop the pieces from ideas already present in the composing score. By and large, this worked well, although the existence of sections laden with revisions indicates that Bach occasionally struggled to arrive at a satisfactory text. Yet in the case of instrumental compositions, especially the keyboard works from the late Leipzig period, for which Bach wrote out fresh fair copies, it may be assumed that he destroyed the composing scores containing sketches and drafts when he no longer needed them. 25 Due to Bach’s process of revision, which led to the creation of new fair copies and rendered early sketches and drafts redundant, keyboard works tend to be preserved in multiple copies. Therefore, Bach studies normally deal with just two types of manuscript: although composing scores contain a considerable number of corrections of a formative nature, the fair copies contain very few or none, and where corrections are found, they are usually limited to rectifying copying errors or slips of the pen. 26 A broader survey of Bach’s autographs indicates that Bach saw in the production of fair copies another opportunity for engagement with the compositional possibilities of the pieces, and continued to revise them as he wrote them out. 27

Sometimes further distinctions are made within these categories. One of these is a revision copy, which lies halfway between a composing score and a fair copy. 28 The score of “Gloria in excelsis Deo” (BWV 191; D-B, Mus. ms. Bach P 1145) is one such example. Evidently working from the score of the Missa (BWV 232I; D-B, Mus. ms. Bach P 180) sometime ca. 1743–6, 29 Bach made multiple minor corrections of a revisionary nature in the score of the “Gloria.” If the appearance of an autograph suggests that it is a revision copy, then it is possible to infer that the music is either a parody or a transcription, even if no such version of the piece is known among the surviving sources. It thus follows that a transcription score is effectively a revision copy. In the score of the harpsichord concertos (D-B, Mus. ms. Bach P 234) traces of Bach’s compositional adjustments can be identified, which are especially numerous within BWV 1052; having carefully studied these, Siegele proposed that it was transcribed from a lost violin concerto. 30 Among the revision copies are also scores in which revisions were worked out at chronologically differing stages. For example, the original manuscript of the D-minor Prelude of WTC II (BWV 875/1: GB-Lbl, Add. MS 35021, f. 4r) was initially produced by Anna Magdalena as a fair copy, which Bach later decided to substantially revise, extending the movement from 53 to 61 bars in length.

Rough and neat are often crucial distinctions when discussing Bach’s handwriting, not only because of the radical differences between the script forms, which may affect the identification of the writer as well as the dating of the handwriting, 31 but also because it reflects how much time and care Bach was able to devote to writing the manuscript. This distinction often goes hand in hand with the compositional stages—the handwriting is rough in composing scores and neat in fair copies. 32 This, however, is not always the case. The calligraphic appearance of the original manuscript of the Prelude in F-sharp minor (BWV 883/1, GB-Lbl, Add. MS. 35021, f. 10r) is quite deceptive; in fact, it contains a number of notational adjustments that Bach had made while producing this score, which strongly suggests that he was not copying from a fully written-out draft, but was more likely writing straight from his head.

Bach’s Own Classification

It is not very clear how Bach classified his scores and parts. The earliest reference to Bach’s musical library is a list of works included in the obituary written by Carl Philipp Emanuel Bach and Johann Friedrich Agricola. The list is, unfortunately, too general to allow any firm conclusions as to the library’s organization. 33 Still, the mention of “five full annual cycles (Jahrgänge) of church pieces, for all the Sundays and holidays,” the first on the list totalling sixteen items, has several important implications. First, it suggests that Bach reused his scores and parts more than a few times during the twenty-seven years of his office in Leipzig; if so, it is reasonable to assume that he had an efficient filing system of all the pieces in the library. Second, the list implies that the original sources of nearly two full annual cycles of cantatas were lost after Bach’s death, as only three annual cycles survive. However, the letter of Johann Elias Bach to Johann Wilhelm Koch, dated 28 January 1741, 34 hints that a large number of cantatas from the two missing annual cycles may have already been removed from Bach’s library in 1740. Such a significant loss of original material has serious implications for Bach scholarship and must be taken into account when interpreting Bach’s creative activities. Whether or not Bach indeed composed all five annual cycles of cantatas will undoubtedly remain a pressing question for future Bach studies.

It is also necessary to consider the origin of manuscripts from Bach’s point of view. Were they the result of the initial composition or were they later fair-copy versions made for the purpose of preservation? Were they intended for Bach’s own library or made for someone else? Regarding the manuscripts that Bach kept in his library, it is likely that he distinguished between two types: (1) original manuscripts, whether complete and incomplete, such as the Orgelbüchein, comprising his personal reference copies; and (2) multiple copies for use by students, or manuscripts that might have been original sources, but were subsequently downgraded to be given away to students or admirers when new versions were created. Awareness of such distinctions is often crucial when evaluating the musical contents of copies made by Bach’s students. For some of his major keyboard works such as WTC, Bach seems to have had multiple copies often containing different versions of the pieces, and his tendency to let his students make copies of earlier versions suggests that he kept his personal copy, containing the most up-to-date shape of the compositions, separate from the multiple copies. The possibility of lost sources cannot be discounted when considering such issues.

As is evident from some of the surviving sources, Bach did not retain all of his manuscripts for his library; some were given away or sold, while others, such as his drafts, have been lost or disposed of. The most important among these will probably have been dedication copies, made with a view of securing a better social status or new employment. The fair-copy score of the Brandenburg Concertos (BWV 1046–1051; D-B, Am. B. 78) and the performance parts of the Missa of 1733 (BWV 232I; D-Dl, Mus. 2405-D-21) are the best-known surviving examples of this kind. Bach would probably also have had fair copies made for private sale. Although no certain sample of this kind exists in Bach’s own hand, a number of copies made by Anna Magdalena, some of which are partial autographs (for example, BWV 1021: D-LEb, Go. S. 3, Faszikel 1), are presumed to have been made for this purpose. 35 Bach is also known to have loaned manuscripts, some of which, such as the original parts of the Sanctus (BWV 232III) sent to Count Franz Anton von Sporck in 1725 or 1726, 36 were never returned, or simply given them away as occasional gifts. Such gifts may in fact be crucial when interpreting how collections such as Bach’s WTC II were compiled and kept together. The source transmission of the F-minor pair (BWV 881) is unique within the collection: its early version was transmitted exclusively in sources connected with Kirnberger, and there are sufficient grounds to speculate that he acquired their autograph while studying with Bach.

Bach must also have distinguished his copies in terms of neatness and general appearance. Those that he considered worthy of sale or presentation to potential patrons or friends, such as the partial autograph score of the Brandenburg Concertos which he dedicated to the Margrave Christian Ludwig of Brandenburg in 1721, could be studied to determine how many errors Bach could tolerate in a presentation copy.

Diplomatic Evidence

The study of manuscripts usually begins with the examination of the document itself using the methodology developed in codicology and diplomatics, scholarly disciplines that critically analyze historical documents by focusing on the conventions, protocols and formulae used by document creators in order to gain an understanding of the process of their creation. In this case, the focus is on the paper, the implements used for writing and erasure and the manner in which the document was crafted and finished. The study of written contents is therefore set aside until the next stage of study.

Paper and Watermarks

It was Spitta who introduced “paper study” to Bach scholarship as an integral part of evaluating the sources. 37 It was an attractive proposition as it offered a new kind of information based on the identification of the approximate time and place of paper manufacture independent of the contents of the document, a study that could be managed entirely objectively. However, as is evident from some of Spitta’s mistakes, the study requires tremendous effort and patience to bear fruit. 38

Until the beginning of the nineteenth century, paper was made by hand in small batches. In Europe, paper makers customarily added watermarks to their paper as trademarks. The designs of watermarks vary widely, some simply comprising the letters indicating the names of the maker (often in initial or in a monogram form), or the region and ruler, while others include elaborate shapes such as coats of arms, plants, animals, tools, etc. The possibilities are endless, but currently no single standard exists to describe each watermark. A common practice in Bach studies is to refer to the catalogue number of Wisso Weiß’s Katalog der Wasserzeichen in Bachs Originalhandschriften (NBA IX/1); 39 other watermark catalogues that can be consulted include the series Papyraceae Historiam Illustrantia published by the Paper Publications Society. 40

Watermarks are created by wireworks mounted in the paper mold. The wireworks (or the paper mold itself, for that matter) did not last very long. The shape of the watermark would have become distorted through repeated use, and when the mold became defective, it was either repaired or replaced. As is often the case with handicrafts, wireworks were never exactly the same, and even minute variations can be detected and evaluated. Because paper makers normally used a pair of molds to increase productivity, another watermark, which may only be slightly different in shape, may reappear in a source in regular sequence.

The chronology of paper production can be pursued only when an adequate sample of watermarks has been gathered. Finding an exact match for a watermark is vital: when the same watermark is found in more than one source, the likelihood that the leaves (folios) containing identical watermarks were manufactured in close proximity, in both time and place, increases significantly. The same inference may then also be extended to the written contents. However, this only allows determining a terminus post quem, that is, the earliest point in time the entries may have been made; the latest possible point, or terminus ante quem, will remain open-ended unless additional evidence becomes available. Bach seldom inscribed dates in his music manuscripts, but because he tended to use the same batch of paper for writing letters and other documents that included dates, it is often possible to narrow down the dating of his music manuscripts to a year or two.

Watermark evidence is not always straightforward to interpret. Research can be impeded if a watermark is partially cut off when the manuscript was trimmed or if it has been concealed within the binding. Watermarks are normally placed in the middle of an uncut, unfolded sheet of paper, and in many instances a sheet was folded and cut before it was used, which would have eliminated or truncated the watermark. 41 Even when watermarks are not physically severed or concealed, they can be difficult to discern if a large amount of ink has been deposited on the paper. Radiographic photography can be used to reveal the images clearly, though very few libraries in the world are equipped to provide this service. Double paper poses an even greater challenge in watermark research. This is an extra-thick sheet of paper that was made by overlaying two sheets before their drying out and designed to prevent ink from bleeding through to the other side while at the same time allowing the writer to scratch out small mistakes with a knife. It may contain two overlaid watermarks, which can be extremely difficult to distinguish due to the paper’s thickness. It is to be hoped that an automatic identification system of digitized watermarks modelled on the Integrated Automated Fingerprint Identification System (IAFIS), or something similar, would become available to researchers in the future. A catalogue of the leaves of Bach’s original sources could then be produced, which would reveal the minute variations in watermarks and help to date the papers of his manuscripts more precisely.

Due to the various sizes and formats in which paper occurs, form, as in studies of other composers, is an essential part of paper studies. Although in Bach studies Weiß’s comprehensive catalogue of watermarks has made this process somewhat redundant, 42 research on non-original manuscript copies on paper not covered by Weiß is yet to be undertaken. Ideally, source studies of secondary copies, including those of works by other composers, should gradually be extended into the same rank as the original manuscripts in order to build up a more comprehensive pool of information concerning the origins of a wide range of papers.

Collation and Binding

The way in which sheets of paper are collated into manuscripts is an essential point for examination. Just like stave ruling (see the “Rastrum and Ruler” in this chapter), it can offer additional clues for a clearer understanding of the writer’s intentions. In the majority of cases, the sheets (Bögen) that form a manuscript are bifolios (double leaves, Doppelblätter), with each sheet folded down the middle to create four sides. 43 Staves are usually ruled after the sheet has been folded. When only two sides are needed—for instance, to insert a newly composed section that would occupy less than two sides, such as the “Et incarnatus” of Bach’s B-minor Mass—a single half-sheet (folio, single leaf) would be used.

Bach’s standard practice, as regards the organization of his music paper (or, more technically, “fascicle structure”), was a binio, 44 or two bifolios stacked up to provide eight usable sides. 45 Departures from this standard practice were probably occasioned by Bach’s circumstances at the time. On the one hand, a typical composing score is a successive gathering of unios, because the composer would not know in advance the exact number of leaves that would be needed. This is the case with Bach’s earliest surviving composing scores (F-Pn MS-2; BWV 134a, performed on 1 January 1719), and became custom practice within a few months upon his appointment as Thomascantor in Leipzig, when he undertook the enormous task of producing a large number of new works. The autograph score of Cantata 105, “Herr, gehe nicht ins Gericht,” for example, is written on five successive bifolios (abbreviated as ‘I × 5’ or ‘5 I’). 46 On the other hand, fair-copy scores of instrumental works from any period often consist of a gathering of quarternios, 47 which is stronger when bound as well as more convenient for handling.

Sometimes the fascicle structure can reveal important facts. The original performance parts of the 1733 Missa, which Bach dedicated to the Saxon elector Friedrich August II on 27 July 1733, were prepared under Bach’s supervision. His family travelled with him to Dresden and assisted him in creating the parts from his score. Carl Philipp Emanuel helped his father with the soprano parts; however, he could not copy all the notes in the planned space of a binio fascicle (eight sides) for the part of Soprano 1, so Bach intervened; he copied from bar 60b to the end using a new bifolio. He ruled twelve systems on its first page, and finished the rest using eleven systems, leaving one unused system below; the three remaining pages were left blank. 48 In practice, this meant that, as a result, the singer would have been holding two physical objects together—a binio fascicle plus a unio—which would have been very awkward in performance, especially at page turns. They would also have had to take great care not to drop the loose leaves! This hypothetical reading of the performance situation is suggested by the present structure of the source (II+I). However, Bach would have probably expected the singer to fold the unio fascicle the other way round and wrap the binio fascicle around it to make a ternio fascicle. This example illustrates the importance of planning the score layout and the collation of sheets together. The fascicle structure thus often reveals the predicted number of leaves required for a manuscript as well as any subsequent adjustments to accommodate new circumstances.

The binding of a manuscript can also expose some fascinating details of the history of its making. An act of affirming a collation of leaves as the final arrangement, it reflects the owner’s expectation for the manuscript to last. A relatively small number of manuscripts were bound by Bach. Known specimens are limited to the keyboard collections and family anthologies such as the Clavierbüchlein for W. F. Bach (US-NH, Music Deposit 31), the same for Anna Magdalena Bach (of both 1722 and 1725, D-B, Mus. ms. Bach P 224 and P 225, respectively), and possibly also Orgelbüchlein (D-B, Mus. ms. Bach P 283), Inventions and Sinfonias (D-B, Mus. ms. Bach P 610), WTC I (D-B, Mus. ms. Bach P 415) 49 and WTC II (the 1744 copy by Altnickol, D-B, Mus. ms. Bach P 430). The last-mentioned volume has a new title page pasted over the original, with an indication that it was to be furnished with a different wording. 50 The change of wording alluded to suggestions that it was Bach who commissioned this copy and requested the change. 51 Indeed, a title page could have been accommodated in the initial plan of collation as fol. 1r of the quarternio fascicle of the fair copy of WTC I as well as of the binio fascicle of WTC II. However, in many instances, title pages were added to the already bound manuscripts at a later stage. The title pages of the four sections of the B-minor Mass (D-B, Mus. ms. Bach P 180), for instance, were once presumably the front parts of the original wrappers 52 for each section; the wrappers would have been cut in half and their (blank?) back parts discarded when the sections were bound in the nineteenth century, possibly during Nägeli’s ownership.

Bindings by later owners, including libraries, can cause difficulties for research. Sometimes the paper has been over-trimmed, with marginal instructions such as movement headers, and even parts of the musical text itself, cut off. Tight binding also makes it extremely difficult to ascertain how a manuscript has been collated. There are also instances where the original collation has been violated, for instance where Auflagebögen 53 (for example, D-B, Mus. ms. Bach P 416, Faszikel 4) have been bound into a volume for the sake of safe keeping, thereby countermanding their practical use.

Rastrum and Ruler

In Bach’s time, music sheets were usually ruled by the writer using a rastrum. Bach, too, would have bought a stock of plain, unruled paper, 54 and prepared his music manuscript paper using an implement with five nibs, drawing one staff at a time. 55 In some instances, he used two implements of different sizes on the same page, accentuating the layout of his score. Two of the best-known examples are the 1736 score of the St. Matthew Passion (BWV 244; D-B, Mus. ms. Bach P 25) and that of the Fifth Brandenburg Concerto (BWV 1050) dedicated to Margrave Christian Ludwig in March 1721 (D-B, Am. B. 78): these show that Bach often had a good plan of the score layout, but had not always executed it perfectly. 56 Occasionally, Bach used a narrower implement to rule additional staves in the bottom margin when he needed to write more notes than he had initially envisaged; sometimes, he also drew a staff line by line with a ruler, or, less frequently, freehand.

In his younger years, Bach tended to draw staves neatly, perhaps using a ruler, a habit that he gradually grew out of. A careful study of Bach’s choice of implement and layout may reveal both his changing attitude to writing scores and the compositional stage at which he arrived having embarked on writing out a manuscript. 57 Instances also exist where Bach had ruled the staves carelessly and unsuccessfully, which prevented him from using the defective area. Perhaps the most famous example is the so-called “unfinished fugue” of The Art of Fugue (BWV 1080/20; D-B, Mus. ms. Bach P 200, Appendix 3, 5). Christoph Wolff argues that Bach stopped writing out this quadruple fugue at bar 239 not because of the reasons C. P. E. Bach cited on the page (“N.B. While working on this fugue, in which the name BACH appears in the countersubject, the author died.” 58 ), but because he was aware of the poor stave ruling in the lower part of the page. Hence from the outset he never intended to write more than he had done, as he had already drafted the so-called “fragment x,” now lost, for the concluding section of the fugue on another sheet. 59

Bach’s rastrum does not seem to have lasted for more than a year, which necessitated him to regularly make or acquire replacements. 60 It would therefore be theoretically possible to reconstruct the chronology of his rastrum usage, very much in the same way as it would be in the study of watermarks: although watermark evidence points to the time of purchase of the paper, rastrum evidence indicates the time the manuscript was prepared. Studying the two sets of chronological evidence together could provide finer details of the activities that took place in the composer’s workshop. For this reason and in consequence of the 1963 publication of an influential article on this subject by Christoph Wolff, the potential value of rastrology is now widely recognized. 61

As no specimen of a rastrum used by Bach survives, research on rastra has to be carried out by measuring the thickness of each line, spaces between them and the overall height of the staves drawn. This actually represents the main impediment for such research: owing to inevitable variations conceivably affected by such factors as the thickness of ink, differences in the skills and habits of the users of the implements, especially the pressure applied to the rastra and angle at which they were held, the surface on which the drawing was done, and so on. It is far from straightforward to work with such a dataset even if the information has been collected thoroughly and comprehensively. Yet by combining this information with other identifiable features, such as the unique defects of each implement (for example, the tendency to fail while drawing specific lines or uneven spacing between the lines) and the unique habits of the user of the rastrum (perhaps the tendency to leave ink blots or bending the implement at the start or end of ruling), future research may be able to arrive at an adequate algorithm to compute distortions in the measurements by taking all these factors into consideration.

Quills, Ink, Pencil and Knife


Although Bach’s quills, like his rastra, have not survived, it is possible to deduce how they were made and used from studying what was written with them. A quill needs to be cut when the tip becomes too soft to write firmly and replaced when further cutting becomes impractical. It is possible to detect the places where Bach sharpened or replaced his quill mid-way through a page, which could offer additional evidence to trace the sequence of his writing. 62 According to Kobayashi, it should be possible to find out both the angle at which the quill was held and the pressure that was applied to it by examining the handwriting using infrared reflectography; the ultimate goal would be to identify the writer. 63 More recently, a statistical approach for writer identification using image processing and machine learning has been proposed by Masahiro Niitsuma. 64


Unlike Bach’s rastra and quills that no longer survive, ink is deposited on paper and can therefore be studied directly. The study of ink is particularly important when distinguishing the layers of revisions as well as identifying subsequent entries made by someone other than the composer.

Ink study has commonly been carried out with the naked eye, but it requires caution. Exposure to oxygen and moisture in the atmosphere over the years affects the chemical composites of iron-gall ink, causing it to fade or change color. In more severe cases, acid contained in the ink can corrode the paper (Tintenfraß). Wide-ranging shades of ink found in eighteenth-century manuscripts can be deceptive and potentially misleading. In fact, a wide range of browns—from very dark, almost black, to light brown—can now be observed on a page on which the writing was originally uniformly black. Thus the evidence gathered by visual impression should be treated with care and, wherever possible, corroborated using scientific, analytical techniques such as spectroscopy and infrared reflectography.

Among the several notable attempts in the recent past to analyze ink using scientific equipment, the first and most rigorously documented was reported by Howard Cox in the mid-1980s. A scientific team in California analyzed the chemical constituents of the ink used in the marginal annotations and the underlinings found in the annotations in the Calov Bible once owned by Bach. Using a method known as PIXE (particle induced X-ray emission), the team concluded that “with high probability, Bach was also responsible for the underlinings and marginal marks.” 65 More recently, Uwe Wolf used a micro X-ray fluorescence analysis technique on the autograph score of the B-minor Mass (D-B, Mus. ms. Bach P 180) to identify and distinguish the inks used by Bach in Leipzig and C. P. E. Bach in Hamburg. Wolf found that the differing amounts of lead contained in the ink they used respectively were a determining factor. 66 Attempts to use spectral analysis of inks on Bach manuscripts, commonly used by forensic document examiners, have produced no significant results. 67


Pencil entries are rarely found in original Bach manuscripts and are not easy to identify with the naked eye because they have faded and are barely visible. A few reported instances of accidentals written in pencil in the original parts suggest that they were added during rehearsal. The best-known example must be Bach’s annotation in the score of The Art of Fugue, an instruction, presumably to his assistant, on how to notate Contrapunctus 8, which Christoph Wolff deciphered using infrared photography. 68 Kobayashi speculates that Bach was compelled to use a pencil as he was bed-ridden at that time. 69 More instances of Bach’s use of pencil may be found in future research using infrared reflectography.


A knife was used both for sharpening the quill and scratching out symbols in need of deletion, which was possible as the paper was usually sufficiently thick and strong. Bach’s knives do not survive, and the matter has, to my knowledge, never been raised as a subject for research in Bach studies.

Notational Evidence

Once the physical and material features of a manuscript have been studied, the investigation turns to the writer of the manuscript and its written contents, especially the notational evidence contained in the source itself. The aim is to broadly assess the circumstances of its production in order to establish who wrote it and when it was written, and to begin to consider for what purpose it was written. All the information gained thus far comes into play for interpreting the musical evidence in this next phase of research.

Bach’s Handwriting—Its Identification and Chronology

Is the manuscript in the hand of Bach or someone else? If it is not Bach, can the scribe be identified? Identification of writers through handwriting analysis is a delicate matter, as graphology is by and large a pseudoscience. Yet with sufficient caution and rigor, it is possible to identify Bach’s handwriting from that of other scribes and establish its detailed chronology.

Nineteenth-century scholars were aware of the potential value of this type of study, but in the process they mistakenly declared more manuscripts to be Bach’s autographs than have subsequently been authenticated. Spitta was aware of the differences between the handwritings of Bach and his second wife, Anna Magdalena, 70 whereas many of his fellow scholars could not tell the difference. Today, it is widely acknowledged that, in addition to Anna Magdalena, several of Bach’s students likewise developed handwriting very similar to that of their master. One of them is Christian Gottlob Meißner (1707–60), 71 a student of Bach’s whose copies were regarded as Bach’s autographs throughout the nineteenth century and beyond. Examining his handwriting, Dürr observed:

Certainly his writings initially show no particular affinity to those of Bach. Only in the course of the years, one symbol after another starts to transform after the model of Bach’s handwriting, and we would have hardly been inclined to recognise the same writer in the earliest and latest shapes, had the ongoing transition of individual forms not been established. The changes, as opposed to the those in the handwriting of the principal copyist A [J. A. Kuhnau], are mostly continuous, less disjointed, and are indeed so numerous that one can recognise, especially in relation to the time he first emerged, a series of characteristic stages of development. 72

Similarly, Anna Magdalena’s music handwriting appears to have gradually transformed in the first couple of years of her marriage to Bach, whose own handwriting went through significant changes during the hectic period of his new career in Leipzig. 73

The present methodology used for identifying Bach’s handwriting and establishing its chronology developed gradually in the first half of the twentieth century, 74 and crystallized in the 1950s in the research of Walter Emery 75 and Georg von Dadelsen. 76 Dadelsen’s monograph of 1958 is a compendium of methodologies for chronological research into Bach’s works. In it, he demonstrates the approaches to the study of Bach’s handwriting, especially its chronology, which in turn plays a central role in establishing the chronology of Bach’s works.

Starting with a detailed study of Bach’s handwriting and using the methodology he had developed, which is still in use today, Dadelsen focused not only on Bach himself, but extended his inquiry to the handwritings of his wife, the two eldest sons and several other copyists from Bach’s circle. 77 The work involved the extraction of a range of specific symbols representative of the writer’s habits and style of writing, such as clefs, minims (half-notes), flagged notes and crotchet (quarter-note) rests, from all the original manuscripts samples known at the time. He then systematically sifted through the samples in order to identify the writer and, at the same time, observe changes in the handwritings. In doing so, he managed to distinguish Bach’s handwriting from those of his copyists, whose handwriting resembled that of their master’s, a task that had previously not been possible.

Having pinpointed the chronological shifts in Bach’s handwriting in the manuscripts that included a date in his own hand, Dadelsen cautiously extended his study to autographs that could be dated according to the circumstances of their production. Having accumulated a sufficient body of reliable evidence, he was able to identify the main features of Bach’s handwritings at various periods of his life. Using this evidence, he further extended the search to autographs that clearly demonstrate the recognizable calligraphic features of each period, 78 although he found a relatively few distinct calligraphic features with which to further distinguish Bach’s handwriting in the manuscripts written between 1724 and ca. 1740. Dadelsen’s work was subsequently improved by Kobayashi, who charted the finer details of Bach’s handwriting from the Arnstadt period to his final years in Leipzig. 79 Sporadic additions of symbols such as accidentals and ornaments continue to be investigated as they have proven to be difficult to identify and analyze. 80

Studying Bach’s handwriting, as well as that of his copyists, involves a careful examination of both Bach’s original manuscripts and those of his copyists (see “Bach’s copyists” later in this chapter). The ultimate aim of such a study is the establishment of the chronology of Bach’s works. Spitta, who could not develop an effective methodology for the study of handwriting, based his inquiries on watermark evidence. Dadelsen, aware that no one had done a thorough and systematic study in this area since Spitta’s time, undertook his chronological study with Bach’s handwriting as the starting point. Around the same time, Alfred Dürr, who had already completed his doctoral research into Bach’s early cantatas, 81 was making significant progress in his study of Bach’s vocal works of the entire Leipzig period, transmitted mostly in the original manuscripts, by focusing on both watermarks (citing the findings by Weiß) and scribes (i.e. Bach’s copyists). 82 The bulk of sources from this period is made up of vocal works, especially sacred cantatas, for which Bach engaged a number of assistants to produce the necessary performance parts. Having examined the numerous copies made by the two principal copyists, 83 Dürr noticed that their handwriting underwent significant changes over the years. Following Dadelsen’s approach, Dürr arrived at the chronology of three annual cycles of cantatas and dated the sources produced in the early Leipzig period; 84 he then mapped out the chronology of the entire Leipzig period (mainly until 1735). Later, Kobayashi applied this methodology to all the original manuscripts, and came up with a new chronology of Bach’s works written between 1736 and 1750. 85 An area in which there is still scope for this type of investigation concerns Bach’s works from his Cöthen period, which could not be dated using diplomatic methods due to insufficient information.

Research into music handwriting has made a significant impact on Bach scholarship in various areas. The identification of a scribe often influences the judgment of the authenticity of a version of a work, a problem faced by the editors of the NBA, as Klaus Hofmann and Yoshitake Kobayashi have outlined. 86 A good example of this is BWV 1050a (D-B, Mus. ms. Bach St 132, Faszikel 1), the evaluation of which has changed after Dürr identified the scribe of the part to be Bach’s son-in-law Johann Christoph Altnickol, 87 and not Gottlob Harrer, Bach’s successor as Thomascantor, as claimed earlier by Heinrich Besseler. 88 Although the identity of the writer has been questioned by Peter Wollny, who identifies him as Johann Christoph Farlau (ca. 1735–after 1770), one of Altnickol’s students in Naumburg, 89 the authenticity of the version, which has to date remained intact, has not yet been challenged. Among the remaining tasks is the investigation of a collection commonly known as the “Great Eighteen” Chorales (BWV 651–668), known to us through a partial autograph (D-B, Mus. ms. Bach P 271, Faszikel 2). The identity of the anonymous copyist and his relationship with the other scribes is crucial for the authentication of the collection as a whole. The first fifteen of the “Great Eighteen” (BWV 651–665) are in Bach’s hand. 90 Of the remaining preludes, two (BWV 666–667) are in the hand of Johann Christoph Altnickol, whose handwriting Wollny has dated to 1751 as terminus ante quem. 91 The final prelude (“Vor deinen Thron tret ich,” BWV 668, survived only fragmentarily), appended at the end of the manuscript, 92 was written by an unknown copyist, known as Anon. Vr or Anon. 12. The first task that Bach entrusted to this scribe was to make a copy of the original manuscript of WTC II around 1742. 93 The copyist carried it out with little respect towards his model, leaving ink marks at page breaks throughout the manuscript. 94 This series of observations supports Kobayashi’s suggestion that the scribe could be identified as Frau Altnickol, Bach’s daughter Elisabeth (“Lieschen”) Juliana Friederica (1726–81). 95 However, the handwriting of Anon. Vr also appears among Bach’s original parts from his final years, 96 and after that, in the copies prepared for C. P. E. Bach in Berlin from the summer of 1749 onwards, 97 which contradicts Kobayashi’s hypothesis, because Lieschen presumably lived in Naumburg with her husband Altnickol, who had been organist there since 30 July 1748. They were married in St. Thomas’s in Leipzig on 20 January 1749. 98 Their son, whom they named Johann Sebastian, was born on 4 October of the same year in Naumburg, but sadly passed away in less than three weeks. 99 No record exists of another child born to the couple until 30 May 1751. 100 Did she really copy manuscripts for both her father in Leipzig and her step-brother in Berlin during her pregnancy? Could the family’s closer ties around the time of her father’s ill health and subsequent death, and the fact that her little brother, Johann Christian, was to be looked after by C. P. E. Bach until 1755, imply that Lieschen offered to help her half-brother as copyist? In addition, because Bach’s collection of music manuscripts appears to have been distributed outside of the legal procedure, the possibility that she, rather than Wilhelm Friedemann, 101 inherited the manuscript cannot be ruled out. 102 If so, C. P. E. Bach obtained the manuscript from her later (i.e., after Altnickol’s death in 1759).

In any case, further evidence is required to argue for or against Kobayashi’s suggestion. A fresh study of Bach’s life in connection with the anecdotal episode involving the “death-bed” chorale, which Bach supposedly dictated, 103 or a study of the reasons for its inclusion at the end of the first edition of The Art of Fugue published in 1751, 104 might uncover such evidence.

Accurate identification of scribes plays an important role in evaluating the contents of a manuscript, especially within the revised passages. Future Bach studies need some form of grapho-analysis capable of examining the subject objectively and scientifically. A recent demonstration by Martin Jarvis, relying on forensic document examination techniques, was an interesting attempt to distinguish between the handwritings of Bach and Anna Magdalena. 105 Although it fell short in its argument that the portion copied by Anna had in fact been “composed” by her, it nevertheless raised renewed hopes of further developments in the improvement of the methodology of grapho-analysis. As Masahiro Niitsuma has recently demonstrated, handwriting analysis in Bach studies will before long be carried out by a computer-driven application that can measure and objectively verify the corpus of handwriting samples. 106

Notational Convention and Chronology

The study of the chronology of handwriting can also inform the chronology of notational practice and vice versa. This is most clearly reflected in the way accidentals were used. Until the beginning of the eighteenth century, the symbol to cancel a pitch raised by a sharp was not a natural, as is customary today, but a flat. According to Dürr, this is a common occurrence in Bach’s autographs written up to 1713, which started to wane in 1714, becoming almost totally extinct after 1715. 107

Another accidental, the older form of which in Bach’s manuscripts over time gave way to modern usage, is the double sharp, an x-shaped symbol raising by a semitone a pitch that has already been raised by a sharp in the key signature. Bach began to change his ordinary sharps to the modern symbol of the double sharp from around 1736. In the autograph fair copy of the St. Matthew Passion (D-B, Mus. ms. Bach P 25) the double sharp is indicated by a large  tilted to the right. 108 It settles into its modern form around 1739, as can be seen in the autograph fair copies of the “London autograph” WTC II (GB-Lbl, Add. MS. 35021). The E-major prelude and fugue of the same set, copied by Anna Magdalena, still used an ordinary sharp to indicate f-double sharp, hinting that Anna’s model—Bach’s autograph score, which may have been Bach’s composing score—predated 1736.

In a similar way, Bach changed his notation of the double flat from an ordinary flat to a larger, plumper symbol, around 1736, as seen in his fair copy of the St. Matthew Passion. 109

Extending the study beyond Bach’s original manuscripts and examining a piece through distant copies, notational practice can sometimes be indicative of a composition’s time of origin, which in turn supports the argument for the work’s authenticity. For example, Bach’s best known organ work, the Toccata in D minor (BWV 565), is only known through copies, the earliest of which—and also the common source for all the other surviving copies—was written by Johannes Ringk (1717–78), a student of Johann Peter Kellner (1705–72), a well-known admirer and collector of Bach’s works (mainly organ music), who claimed to have met Bach. 110 Ringk’s copy (D-B, Mus. ms. Bach P 595, Faszikel 8) uses a modal key signature and 3-shaped flags on semiquavers (sixteenth-notes), notational conventions that had petered out by the middle of the eighteenth century. 111

Bach’s Copyists

The identification of Bach’s copyists (i.e., those who participated in the production of the original manuscripts) and an understanding of the extent of their involvement and quality of work is an important part of research, as it paints a vivid picture of the actual practical situations in which the manuscripts were made and used. It also provides crucial evidence for chronology research, as already discussed within the section “Bach’s Handwriting—Its Identification and Chronology,” while an insight into a copyist’s habitual errors would offer valuable clues for textual analysis.

The copyists involved in the production of Bach’s Leipzig cantatas have been particularly well researched. The scores—often composing scores, especially in the case of the early Leipzig cantatas—are usually in Bach’s hand, whereas the performance parts were entrusted to his copyists, although Bach may subsequently have made small corrections and added performance-related marks. Bach’s copyists were normally drawn from the ranks of the pupils at the St. Thomas School, family members and private students. Dürr’s main contribution to this aspect of research was the identification of their roles and distribution of work, classified into two types: principal copyists (Hauptkopisten), to whom Bach entrusted the preparation of entire sets of parts (without duplicates) from scores, 112 and other copyists, who made duplicate parts. The most important of Bach’s principal copyists were: Johann Andreas Kuhnau (= Hauptkopist A), C. G. Meißner (= Hauptkopist B), and Johann Heinrich Bach (= Hauptkopist C) from the early Leipzig period; Johann Ludwig Krebs and Johann Ludwig Dietel (= Hauptkopist F) from the middle Leipzig period; and Johann Nathanael Bammler (= Hauptkopist H) from the late Leipzig period. 113 Some of his family members—Anna Magdalena, Wilhelm Friedemann, Carl Philipp Emanuel, Johann Christoph Friedrich and Johann Christian—also served Bach faithfully as his copyists. Although a number of anonymous copyists have been named since the publication of the second revised edition of Dürr’s seminal work in 1976, 114 a comprehensive list of scribes is yet to be assembled. Their names and work were updated and summarized in a catalogue of Bach’s copyists by Kobayashi and Beißwenger, published in 2007 (NBA IX/3). 115 Statistical data, for example, on the copyists’ habitual errors—essential for evaluating the quality of text in the copies—are some of the issues that remain to be investigated in future studies.

Other Scribes of Bach’s Works

The previously mentioned catalogue of Bach’s copyists does not include information about the non-original manuscript copies of Bach’s works that were not made for the composer’s own use, such as the copies of keyboard music his students created as part of their studies with him. While a comprehensive catalogue of all the scribes and their copies is yet to be compiled, 116 there are some notable studies on key scribes. The most important scribe from Bach’s youth is his eldest brother, Johann Christoph (1671–1721) of Ohrdruf, whose hand appears together with Bach’s in two well-known anthologies—the so-called “Möller manuscript” (D-B, Mus. ms. 40644) and the “Andreas Bach Book” (D-LEm, III.8.4), which are believed to have been copied between 1705 and 1713. 117 These are important sources for understanding Bach’s creative development, which is explored in Section IV of this volume. Among the best known scribes of Bach’s works from his immediate circle are: Johann Gottfried Walther (1684–1748), 118 Johann Tobias Krebs (1690–1762) 119 and Johann Caspar Vogler (1696–1763) 120 from the Weimar period; Heinrich Nikolaus Gerber (1702–75) 121 and Bernhard Christian Kayser (1705–58) 122 from the early Leipzig period; and Johann Friedrich Agricola (1720–74) 123 and Johann Christoph Altnickol (1719–59) from the later Leipzig period. Johann Christian Kittel (1732–1809) 124 and Christian Friedrich Penzel (1737–1801), 125 pupils of Bach in the final years of his life, continued to copy his works after his death, and promoted their master’s works in Erfurt and Merseburg, respectively.

Whereas much of the research on copies made after Bach’s lifetime belongs to the study of the reception of Bach’s works, a number of them attest to the early shapes of the compositions, thus providing an additional insight into Bach’s creative process. Notable examples include the manuscript score of the early version of the St. Matthew Passion (BWV 244b; D-B, Am. B. 6 and Am. B. 7), published in facsimile in NBA II/5a as a copy in the hand of Johann Christoph Altnickol, 126 but which, according to Peter Wollny, was copied by Altnickol’s student Farlau. 127 Other examples include Johann Christoph Oley’s (1738–89) copy of the Italian Concerto (BWV 971), 128 and Christoph Ernst Abraham Albrecht Freiherr von Boineburg-Lengsfeld’s (1752–1840) copies of the WTC II and Goldberg Variations. 129

Musical Evidence

Having established the diplomatic and notational evidence, evaluation of the writer’s activity can proceed through an examination of musical evidence manifested in the form of written contents. This needs to be approached from the actual context of the manuscript production by piecing together the evidence already gathered, and to take into consideration a range of issues. These range from the manner in which the manuscript was prepared and its purpose, to the actual act of committing the notes to paper, or the way in which they were written and revised, with the ultimate aim of reconstructing the sequence of events to reveal the intentions of the writer. The study of the composition and revision processes is absolutely central in manuscript studies, as it harbors a compelling story about the composer and his work, including the context of its creation and revision, 130 its history built on factual evidence. However, one should be aware of potential pitfalls when attempting to interpret information flowing from the composer’s brain to his hand, and on to paper, as the written contents cannot convey everything the composer imagined. For instance, some ideas will inevitably be lost when they are committed to paper (although, of course, there is no way to prove it), while others may not be fully formulated; 131 some may have been written inaccurately, 132 or incompletely, 133 and in some the errors may have been rectified when the composer had subsequent opportunities to check or write the piece out afresh. 134 A good example is the way in which Bach treated articulation and phrasing marks in his scores: he seldom fully wrote them out. Occasionally, however, Bach does the opposite and supplies “unexpected extra” instructions that cannot be explained with logic or consistency. 135 All this may be appreciated from a broader historical context according to which in Bach’s time notated music was not expected to represent a work in every detail, but was merely a kind of template into which performers were expected to inject “life” by providing articulations, embellishments, phrasing, and even improvisation.

In addition to the study of the processes of composition and revision, manuscript studies offer further avenues to gain new knowledge from the sources with regard to performance implications, Bach’s teaching, critical editions, and, finally, transmission and reception history.

Performance Implications

Bach’s original manuscripts reveal information from which conclusions can be drawn with regard to the way in which he expected his works to be performed in terms of his performing forces, articulation, phrasing and possibly other aspects. Arguably the most intensely debated, as well as the most consequential for the performance of Bach’s vocal works in the past three decades, is Joshua Rifkin’s revolutionary idea that Bach’s vocal works were routinely sung by one singer per part, 136 which contradicts the standard and traditional interpretations (e.g., three singers per part). Rifkin claims that this is manifested in the surviving original parts, although from the viewpoint of source study, his assertion lacks conviction, and has consequently attracted numerous objections from other scholars. 137 The argument is complex, as source evidence is sometimes incomplete and ambiguous. For example, the absence from the surviving vocal parts of indications that would specify which movements are to be sung by the concertists alone and which are to be joined by ripienists may be interpreted as a specific instruction for not using ripienists in the piece, or perhaps as Bach’s desire to maintain flexibility, which would allow him to reuse the parts, possibly in different settings, in the future. Possible deformations of source evidence, for instance, the disposal by subsequent owners of duplicate parts that were once kept together, must also be considered. 138 Each piece of evidence must therefore be interpreted within its own broader historical context, which in itself is in need of clarification and further research. 139

The performance markings Bach himself added to the original manuscripts, such as ornaments, slurs and dots, have been reasonably well-researched 140 and understood, and it has become standard procedure among the circles of early music performers today to examine the original sources in order to understand Bach’s approach to performance issues. Particularly notable is the shift in recent decades from scores that are often available in facsimile editions to the reproductions of original performance parts that have not been easily obtainable. This is necessary because Bach added to the individual parts performance-related marks that are not indicated in his original scores, and is therefore another area requiring further research. 141

The high visibility of such issues does not mean that everything is clearly understood and that no room is left for further study. Many questions remain unresolved. One such question concerns the placement of slurs—how many notes are covered by a single stroke: even in some of the most exquisite autograph fair copies, such as that of the Violin Solos (BWV 1001–1006; D-B, Mus. ms. Bach P 967), this is not always clear. 142 When it comes to works for which Bach’s autograph scores do not survive, such as the Cello Suites (BWV 1007–1012), we have to rely on copies in which the accuracy of a slur’s placement is often highly questionable. 143 In a set of parts made by Bach and his copyists, one may find that the placement of a slur varies from one part to the next, thus posing a problem for the editor of a critical edition. 144 It seems that Bach was aware of the inherent problem of the notational system in this regard. From around 1746, he began to use dots as articulation marks to indicate the notes that did not belong with the slurred notes, probably to overcome this problem; this can be seen in the performance parts of BWV 8 (later version in D; Bach-Archiv Leipzig, Thomana 8), in which the principle is followed through with particular meticulousness in the traverso part. 145 The dots found in Altnickol’s 1744 copy of WTC II, as later additions to the subjects of the A-minor and B-flat-minor fugues, probably date from the same period. Given the new role the dots assumed around that time, caution is required when interpreting their meaning. 146

The subtleties in Bach’s beaming of quavers (eighth-notes) suggest that his phrasing and articulation indications went beyond the use of symbols. 147 Following the convention of his time, Bach normally wrote beams for a succession of quavers longer than the unit of a beat: for example, one beam was applied for six quavers in 3/4 and two for eight quavers in 4/4 . It is rare to find Bach consistently writing quaver beams over a one-beat unit (e.g., in 3/4: ). When he does so, his choice of notation seems to imply that he wanted to stress a specific mood or the character of the movement. Study of these exceptional cases has revealed that these were used to indicate two types of situations: (1) a lively execution of leaping quavers; as in WTC I, Fugue in C-sharp major (BWV 848/2); Prelude in F-sharp minor (BWV 859/1); Prelude in G major (BWV 860/1); Prelude in B-flat major (BWV 866/1); Variations 28 and 29 of the Goldberg Variations (BWV 988): and (2) a plodding mood in quavers of the same pitch or in step-wise motion, as in WTC I, Prelude in F minor (BWV 857/1), Prelude in G minor (BWV 861/1); WTC II, Prelude in F minor (BWV 881/1), Prelude in A minor (BWV 889/1); Orgelbüchlein, “Da Jesus an dem Kreuze stund”( BWV 621), “Es ist das Heil uns kommen her” (BWV 638), “Ich ruf zu dir, Herr Jesu Christ” (BWV 639); and Invention in B minor (BWV 786). In the majority of movements, however, Bach uses both types of quavers discreetly to convey extra nuances, be it to distinguish two contrasting motifs, to articulate structurally at cadential passages with default beams, or to indicate the increase of attention to the local passage with extended beams. Bach’s quaver beams may be a new resource for finding out how Bach responded to his pieces in performance. Although the process of writing music cannot be translated directly into how the writer actually performed the piece, we may have come one step closer to learning how Bach engaged with his music, or at least acquired an additional means to read Bach’s intentions as to how he perceived his music at multiple layers of compositional thought process from motivic to structural levels. 148

Bach’s Teaching

Some of Bach’s manuscripts, particularly those for keyboard works, document how and what he taught his students, among whom were his own sons and wife. As has been demonstrated by numerous scholars since the time of Spitta, considerable insights into Bach’s pedagogical methods and educational philosophy can be gained by careful study. 149 Perhaps the most important source that epitomizes Bach’s systematic teaching program is the Clavierbüchlein for Wilhelm Friedemann Bach (Yale University Library, New Haven, Music Deposit 31), a notebook he gave in January 1720 to his then nine-year-old son. The manner in which the pieces were entered into the notebook reveals how Bach taught this gifted child in the course of several years. Comparison with later collections that evolved from this notebook, namely WTC I and the Inventions and Sinfonias, shows that Bach adapted and improved his earlier teaching materials as tools for teaching advanced students. In addition to these two well-known collections, there are other collections of keyboard works—the Six Small Preludes (BWV 933–938), English Suites and French Suites, as well as those that appear in C. P. E. Bach’s Nachlass, 150 namely “5 Präludien und 5 Fugen,” 151 “Suite pour le Clavecin” 152 and “Einige Clavierstücke und Fugen von Joh. Seb. und Wilhelm Friedem. Bach.” 153 —that require further study so that a broader view of Bach’s pedagogy can be attained, or even just to establish whether some of them are indeed genuine products of Bach.

Occasionally, copies made by Bach’s students contain sporadic additions of performance-related marks such as slurs, ornaments and fingerings, as manifested, for example, in Johann Caspar Vogler’s copy of the early version of the opening prelude of WTC II (BWV 870a/1, D-B, Mus. ms. Bach P 1089, f. 4v). 154 Because these symbols do not contain sufficient calligraphic information to connect them to Bach, it is necessary for these additions and their musical context to be studied. 155

Critical Editions

Most of Bach’s works are known to us through manuscripts, some of which are autograph or original manuscripts (e.g., performance parts with autograph revisions), while others survive only in copies. The quality and reliability of text transmitted in manuscript copies vary. As a result of this varied nature of manuscripts arose the idea of a critical edition, an attempt to produce a definitive engraved version of the music, accompanied by a critical commentary that explains the reasons behind the choice of a particular version or versions and lists alternative readings. Such editions are one of the primary means by which the work of musicologists reaches performers.

One of the first tasks for an editor is to identify the most reliable source among the surviving sources by carefully assessing them (recensio) to establish their common readings and variants, together with the origins and chronology of these readings. This is described in a tree diagram (stemma), 156 which elucidates the relationship of sources (filiation). The primary sources, containing the composer’s original version(s) of text, 157 are distinguished from the rest, the secondary sources. This can be a very lengthy process because evidence is often both extensive and complex. Separating Bach’s revisions from alterations by others (and identifying the hand where someone else entered the revisions/alterations), distinguishing the layers of revisions Bach entered in a single manuscript or identifying the number of copies in which Bach entered his revisions, including those that are no longer extant, are among the most challenging tasks. Ideally, answers should be drawn using philological methodologies, which include studying the patterns of errors and variants reflected in subsequent generations of copies, and combined with the study of ink 158 and calligraphy. 159 However, due to a lack of evidence, our findings are often conditioned by our received understanding of Bach’s compositional process. (On the other hand, the same understanding may also color our perceptions and make us biased.)

The notion of versions and variant readings (or variants) as well as their differences is important. Both concern situations where the musical text of the same piece is handed down to us through the surviving sources in non-unanimous form. A version is a difference observed at a large scale, such as the length, form or textural shape of the composition, whereas a variant is a smaller-scale discrepancy manifested among the surviving sources that belong to the same version of the composition, in features such as pitch and rhythm. 160 A new version may have been produced when Bach had fresh motivation. Such an occurrence, however, may never be ratified, and can therefore not be considered in the production of an edition. One of the most fundamental tasks of an editor is to assess all the differences found in the sources and determine whether they make up multiple versions or are merely variants. Ideally, critical editions should produce different versions in full, while variants should be dealt with in a critical report in list or table form, with the more important variants considered as alternative readings and supplied to the respective musical pages as footnotes.

Under the banner of a critical edition, mixing variants from different versions, called conflation of texts (contaminatio), is considered a violation of editorial rules, as it creates a new “hybrid” version of the text, which the composer probably never considered. 161 Producing a single version of the text from a pool of multiple versions therefore requires great caution and sufficient justification for the choice of one version over another. Once these necessary steps have been completed, the next phase, aimed at establishing whether the musical texts of the selected primary sources were really intended by the composer (examinatio), can then proceed. Contrary to common belief, Bach did make occasional mistakes, ranging from simple grammatical errors, such as parallel fifths and octaves or inaccurately notated rhythm of a complex constellation of beamed notes (lacking a beam, for instance) to notational oversights such as missing or inaccurate accidentals caused, for instance, by a confusion between modal and modern key signatures 162 or in places in which a piece would temporarily modulate. 163 Occasionally, the received understanding of Bach’s compositional logic and procedure influences text-critical judgment, and results in the correction of a grammatically valid pitch. This tendency was particularly strong in the editorial decisions of the NBA.

Bach scholarship has greatly benefited from the NBA project, which, in addition to producing definitive versions of Bach’s music for further study, has provided scholars with an array of valuable insights acquired during the process of its making, as has already been touched upon. However, the process has not finished with its completion in 2008. 164 Volumes of the revised edition of NBA (NBArev) have been published since 2010 to replace those that needed to be re-edited as a result of advancements made since the time of their original publication both in terms of the concept of a critical edition and the state of scholarship. In addition, a number of hitherto unknown sources and those once considered lost have resurfaced. The mounting successes of scholars and their promotion of scholarship have resulted in a growing desire among the general public to hear unusual versions of well-known pieces, for example, an early version of the St. Matthew Passion (BWV 244b). The need for further research seems to continue without end as research reveals new insights. As our understanding of Bach’s music continues to develop, our notion of what a critical edition should be will also change. Scholars in the future will undoubtedly discover new ways of presenting the complex yet fascinating continuum in the transformation and maturing of a work in the hands of its composer, which the sources harbor, but which tends to get buried in the dense, utilitarian and often not easily digestible make-up of critical reports. 165

Transmission and Reception History

Transmission (Überlieferung) is a term used in manuscript studies to describe the way in which works were disseminated through the act of copying, as this was the only practical means available at the time. Just as the biological features of living creatures are passed on to their offspring, each of which, however, is born into this world different from the other, so too do music manuscripts, which, although fundamentally retaining nearly all the genetic information from their parental source, undergo modification in the process of copying. The reasons behind the newly introduced changes are numerous, and unmasking them is one of the primary objectives for establishing a stemma. 166 A lack of competence on the writer’s part to accurately reproduce (or properly update the notational convention of) the model or a change of score layout necessitated by the size of notation or paper, are among those reasons. Variants could also have been introduced deliberately, where the writer wished to correct or modernize some of the notational or musical properties of the written contents. The modifications are then evaluated as either errors or variants, depending on the musical context. For example, pitch variants will always be examined within the context of both the generic harmonic grammar and specific stylistic traits, such as the motives and figures used within the piece.

A careful evaluation of the musical text is sometimes extended to cover the people involved, copyists, users and owners. When the subject touches upon the biographical sphere, the study, exploring the musical lives of individuals and their engagement with the works in question within the contexts of a particular time and region, becomes one of reception history. A notable example from recent scholarship that successfully illuminates how tracing the sources’ transmission can contribute to our better understanding of the reception history is a study of a group of recently resurfaced manuscript sources of Bach’s Sonatas and Partitas for Solo Violin (BWV 1001–1006): it examines their provenance, some of which are linked to eminent figures such as Carl Friedrich Zelter (1758–1832) and Louis Spohr (1784–1859), establishes a textual link between them, traces the origin of this branch of transmission to Bach’s autograph and manages to paint a lively picture of how this celebrated work of Bach’s had been valued by many. 167 The nature of reception study often requires scholars to branch out into spheres of research that lie outside their main subject area (i.e., spheres that focus on the individuals and situations to which a particular source is linked), so its future lies in the collaboration between scholars with different specialist backgrounds. The knowledge gained through the exploration of specific people, periods and regions should then be reassessed in context of the placement of its subject within broader current of musical scenes and trends.

Future Work

The breadth and depth of detail at which each scholar engages with Bach manuscripts as discussed in this chapter gives the impression that it is impossible for any one scholar to be fully cognizant of every new development in research as well as the past achievements. Yet it is imperative that we are aware of the complexity of the subject. Even when engaging with a specific field of manuscript study, crucial evidence often comes from a holistic and systematic investigation, as was repeatedly demonstrated by scholars in the past. As research grows ever more specialized and segmented, the future calls for a research environment in which scholars will share work and knowledge for their mutual benefit, while focusing on their particular research interests. Only in this way shall we avoid losing touch with one another’s achievements. As I suggested in 2002, Bach scholars need to consider creating a flexible, stable and dynamic global research infrastructure for sharing data, computing, and other resources via the internet. Using the recently established e-Science model is, in my view, the most attractive option. 168

The course of study is affected, often fundamentally, by new source finds. The discovery of Bach’s personal copy of the Goldberg-Variations in 1974, which included a hitherto-unknown set of Fourteen Canons (BWV 1087) 169 in his hand at the back of the volume, provided rich biographical material to repaint the picture of Bach in the last few years of his life. 170 The rediscovery by Christoph Wolff and his colleagues at the Ukrainian Research Institute, Harvard University, of the Berlin Sing-Akademie collection in Kiev in 1999 likewise shook the foundation of C. P. E. Bach studies, which subsequently became the main fuel for Carl Philipp Emanuel Bach: The Complete Works, the volumes of which started to appear from 2005 at an astonishing pace. 171 Wolff’s next project was Expedition Bach at the Bach-Archiv Leipzig. Since its inception in 2002, his team has been systematically searching for historical documents in federal, communal and church archives in the historical territories of Saxony, Thuringia and Anhalt, and has already reported a number of significant finds including a hitherto-unknown autograph of the Weimar aria “Alles mit Gott und nichts ohn’ ihn” (BWV 1127) in 2005, 172 and, in the following year, the so-called “Weimar Organ Tabulature” bearing Bach’s own signature. Having been widely accepted as the earliest manuscript copy in Bach’s hand, and containing the organ music of Reincken and Buxtehude, the Tabulature provides us with concrete material to reassess the ability of the teenage Bach, the young organ scholar. 173 Naturally, the net ought to be cast wider in the future. The source hunt in the archives in Central and Eastern Europe, as outlined by Szymon Paczkowski in his keynote paper entitled “Bach and Poland in the 18th century,” read at the Bach Network UK dialogue meeting in Warsaw in July 2013, 174 will soon be launched with the hope of uncovering more evidence relating to the early reception of Bach’s works by Kirnberger in Częstochowa, Podhorce, Równe and Lwów, and Mizler in Końskie and Warsaw. Further discoveries of Bach’s original manuscripts are anticipated in the future from pursuing these teacher-pupil connections. 175 Meanwhile, revisiting already known collections and archives has proven to be useful. Among the collections housed in the Staatsarchiv Leipzig, a branch of the Sachsische Staatsarchiv, are the sources that have been transferred from the company archives of music publishers such as Breitkopf & Härtel and C. F. Peters, which were for some unknown reason overlooked by the editors of NBA, despite having been long available to scholars. 176 It was only in 2013 that the vast corpus of Bach sources housed there came to be known, and the process of reevaluation began. Among the first fruits of this fresh investigation was the recovery by Christine Blanken of Breitkopf’s house copies, which include a very early fair copy of the keyboard Toccatas in the hand of a Bach pupil and containing Bach’s autograph revisions, many early copies of the compositions of the Bach family, J. S. Bach’s organ chorales copied by C. G. Gerlach and J. L. Krebs and copies of Bach’s organ works published by A. B. Marx in the 1830s. 177 Without doubt, further progress will be made in the near future, and the report is keenly awaited by many scholars. Alongside the recovery of physical sources, we will learn more about the lost sources by piecing together various fragments of information arrived at through a systematic investigation of known sources. For example, the systematic research of Nobuaki Ebata on manuscript sources of the Four-Part Chorales may well lead to the identification of some that originated from the chorale movements of the lost Picander-Jahrgang cantatas (the fourth annual cycle, of which presently only approximately ten cantatas are known to have survived), and possibly offer further possibilities of identifying the chorale movements from the cantatas of the fifth Jahrgang and two passions, all of which have completely disappeared. 178

Equally important is the establishment of the e-Science environment with new methodologies and research techniques using Artificial Intelligence techniques to extract more data from the existing sources and to carry out more powerful analyses. The new analyses may yield connections impossible to make through our cognitive power alone. For example, they will hopefully show the relationships between all the variants found in sources of a single work and how to measure the strength of these relationships, with all that this implies for their provenance. In this light, nearly every area of source study needs to be revisited to ensure that every new research resource, knowledge or analytical technique will be integrated dynamically into the system, so that it will be available to everyone in real time.


I am most grateful to Nobuaki Ebata, David Ledbetter and Richard Rastall, whose comments on an early version of this chapter were invaluable. I am also indebted to Tanja Kovačević and Robin A. Leaver for their editorial work on this chapter.

TBSt 1 and TBSt 4/5.

Dürr Chr 2. The work of Dadelsen and Dürr was further refined by Yoshitake Kobayashi; see Kobayashi, and Yoshitake Kobayashi, “Quellenkundliche Überlegungen zur Chronologie der Weimarer Vokalwerke Bachs,” Karl Heller and Hans-Joachim Schulze, eds., Das Frühwerk Johann Sebastian Bachs (Cologne: Studio, 1995), 290–310; and NBA IX/1 (1985), IX/2 (1989), and IX/3 (2007).

Concise summaries of the main sources of vocal works that were known by the late 1980s are found in BC. A brief list of all the sources of Bach’s entire oeuvre is found in BWV2 (1990); one must, however, exercise caution when using it, as it merely lists the sources, without applying rigorous source-critical considerations.

Particularly important is the series of facsimiles issued by Insel Verlag in the 1920s. Made before ink corrosion had damaged them even further, they preserve the clearest images of Bach’s autographs. Another representative collection of Bach’s autographs was published in a series called “Faksimile-Reihe Bachscher Werke und Schriftstücke” (1954–88), and its successor “Faksimile-Reihe Bachscher Werke und Schriftstücke. Neue Folge” (2001– ). Anyone using a facsimile edition should be alert to any errors or imperfections that might have crept in during its preparation and production. See, for example, Albi Rosenthal, “Facsimiles as Sources of Error,” Paul Brainard and Ray Robinson, eds., A Bach Tribute: Essays in Honor of William H. Scheide (Kassel: Bärenreiter, 1993), 205–207.

The most important is Die Bach-Sammlung: Musikhandschriften der Staatsbibliothek zu Berlin – Preußischer Kulturbesitz, Teil 1 (Munich: Saur, 1997–2003; Supplement [I] (2000), and Supplement II, 2003).

Some of the most prominent at present are: Staatsbibliothek zu Berlin (, Bayerische Staatsbibliothek ( and Sächsische Landesbibliothek – Staats- und Universitätsbibliothek Dresden ( See also note 5.

For a discussion of this issue, see Yo Tomita “Challenging Virtuality: A Response to Ruth Tatlow,” Bach Notes: The Newsletter of the American Bach Society 18 (Spring 2013): 3–6.

For example, the term “discarded autograph” represents one of the five categories of sources used in Chopin studies. It refers to a manuscript that Chopin initially intended as a clean copy for the engraver, but as he subsequently revised the score, the copy, no longer suitable for its original purpose, was discarded.

See further “Critical Editions” in this chapter.

Distinction is also made between a corrected “personal copy” (Handexemplar) and a copy in which corrections were transferred from an authentic source by another hand. In many cases, Handexemplars survive while the original manuscripts themselves do not. See Butler in Chapter 2.

See Nobuaki Ebata, Yo Tomita and Ian Mills, “Mendelssohn and The Schübler Chorales (BWV 645–650): A New Source found in the Riemenschneider Bach Institute Collection,” BACH 44/1 (2013): 1–45.

See the critical review of NBA II/1 (1954) and NBA II/1 KB (1956), by Georg von Dadelsen: Mf 12 (1959): 315–34; ET “Friedrich Smend’s Edition of the B-minor Mass by J. S. Bach,” BACH 20/2 (Summer 1989): 49–74.

Although both terms basically mean the same thing, “autograph” is generally the preferred term in manuscript studies. For a debate on the terms, see (accessed 26 February 2014).

See Alfred Dürr’s humorous article “De vita cum imperfectis,” Robert L. Marshall, ed. Studies in Renaissance and Baroque Music in Honor of Arthur Mendel (Kassel: Bärenreiter, 1974), 243–53.

Wollny identifies the scribe as Johann Lorenz Bach (1695–1773), grandson of an elder brother of J. S. Bach’s father. Peter Wollny, “Einführung” / “Introduction,” Johann Sebastian Bach: ‘Nun komm, der Heiden Heiland’ BWV 61. Faksimile der Originalpartitur (Laaber: Laaber, 2000), vii and xii–xiii.

See NBR 204 (No. 208) and Dok 2: 338–9 (No. 439) for the account recorded by a clerk of the Leipzig town council dated 17 March 1739. See also Kobayashi, 44.

Peter Wollny, “Neue Bach-Funde,” BJ 83 (1997): 7–50, esp. 42–3; Peter Wollny, “Vorwort,” Peter Wollny, ed., Johannespassion: Passio secundum Joannem, Fassung IV (1749), BWV 245/BC D 2d, mit der unvollendeten Revision BC D 23 (1739) im Anhang, für Soli (SATBB), Chor (SATB), 2 Traversflöten, 2 Oboen/Oboe d’amore, Oboen da caccia, Viola da gamba, 2 Violinen, Viola und Basso continuo / Johann Sebastian Bach (Stuttgart: Carus, 2002), vii and ix.

See James B. Welch, “J. S. Bach’s Concerto in D minor, BWV 596, after Vivaldi: its Origin, Questioned Authorship, and Transcription,” Diapason 74 (May 1983): 6–7. For further discussion of Bach’s transcriptions of works by his fellow composers, see Chapter 10.

“London original manuscript” (Londoner Originalhandschrift, as used in NBA V/6.2) would be a more precise designation than this English nickname introduced by Richardson and Emery in their respective articles published in Music & Letters and used widely since at least 1953.

See Robert L. Marshall, The Compositional Process of J. S. Bach. A Study of the Autograph Scores of the Vocal Works, 2 vols. (Princeton: Princeton University Press, 1972), 1: 3–6 and 31–32, for definitions.

Marshall, Compositional Process, 2 transcribes all the known sketches and drafts. NBA Supplement (Wollny and Maul, 2012) reproduces them in facsimile. The nature of the discarded readings is wide-ranging. One of the most fascinating examples includes the draft of BWV 149, the paper of which was used for the score of BWV 201 (D-B, Mus. ms. Bach P 175) several weeks later.

See, for example, Yo Tomita, “Bach and His Early Drafts: Some Observations on Little Known Early Versions of Well-Tempered Clavier II and the Goldberg Variations from the Manfred Gorke Collection,” BACH 30/2 (1999): 49–72.

Dadelsen TBSt 4/5: 58–72; Marshall, Compositional Process, 1: 4.

This observation is particularly relevant to the manuscripts of solo keyboard works. There are only a few fair-copy scores of his ensemble music, as Bach presumably considered the production of parts to be the equivalent of making a fair copy.

Marshall, Compositional Process, 1: 5, 18–20.

This dating is by Kobayashi, 52. For a long time, the year 1745, as proposed by Butler, has been known as the possible date of performance, but recent studies by Rathey and Leaver propose that the revision copy of the “Gloria” originated a few years earlier, namely in 1742. Gregory Butler, “Johann Sebastian Bachs Gloria in excelsis Deo BWV 191: Musik für ein Leipziger Dankfest,” BJ 78 (1992): 65–71; Markus Rathey, “Zur Entstehungsgeschichte von Bachs Universitätsmusik ‘Gloria in Excelsis Deo’ BWV 191,” BJ 99 (2013): 319–28; and Robin A. Leaver, “Bachs lateinische Kantate ‘Gloria in Excelsis Deo’ BWV 191 und eine lateinische Rede über Lukas 2:14,” BJ 99 (2013): 329–34. However, Rathey and Leaver’s new dating is not without problems. A careful consideration of the evidence presented by Kobayashi and his dating “von etwa 1743 bis etwa 1746,” especially in conjunction with watermark evidence (Weiß 21) and Bach’s G-clef, renders 1742 to be too early. Future research must include a reassessment of watermarks to see whether or not the paper is a variant of Weiß 21.

Ulrich Siegele, Kompositionsweise und Bearbeitungstechnik in der Instrumentalmusik Johann Sebastian Bachs (Stuttgart: Hänssler, 1975), esp. 101–16; see also NBA KB VII/7 (1971): 36–40. The lost but reconstructed concertos were published as NBA VII/7 (Supplement), and BWV2 (1990) indicated them by R-numbers, such as BWV 1052R. R-numbers were abandoned in BWV2a (1998).

Dadelsen uses the terms working script (Gebrauchsschrift) and calligraphic script; see TBSt 4/5: 71–2; see also Marshall, Compositional Process, 1: 4.

See, for example, Bach’s autograph score of the Christmas Oratorio (D-B, Mus. ms. Bach P 32), which includes many parody movements from secular cantatas.

NBR 304 (No, 306); Dok 3: 86 (No. 666).

NBR 210 (No. 219); Dok 2: 388 (No. 484).

See Yo Tomita, “Anna Magdalena as Bach’s Copyist,” UB 2 (2007): 59–76, esp. 71.

Or more precisely, after Christmas 1724 and before Easter 1727; see Dürr Chr 2, 77 and 93.

Spitta is enthusiastic in relating his theory of using watermarks in Bach source studies; Spitta 2: 775–776; Spitta ET 2: 680.

Even though Spitta believed his method was vigorous and thorough, conclusions he drew from incomplete and imperfect data led him astray; for instance, he failed to recognize some of the crucial features of watermarks, such as the three variants of the watermark “MA” (small, middle and large). Awareness of such features allowed Wisso Weiß to classify and date Bach’s paper more precisely than Spitta. See Dürr Chr 2, 35, 126, 138–141.


Volumes that are most commonly referred to in Bach scholarship are: Edward Heawood, ed., Watermarks: mainly of the 17th and 18th centuries. Monumenta chartae papyraceae historiam illustrantia, I (Hilversum: Paper Publication Society, 1957); Émile Joseph Labarre, ed., The Nostitz papers: Notes on Watermarks found in the German Imperial Archives of the 17th & 18th centuries, and Essays Showing the Evolution of a Number of Watermarks. Monumenta chartae papyraceae historiam illustrantia, V (Hilversum: Paper Publication Society, 1956); see also the database Musik der Dresdner Hofkapelle. Schrank II. Die Instrumentalmusik zur Zeit der sächsisch-polnischen Union. Wasserzeichenkataloge.

See note 45.

Weiß 1: 14, for a general discussion of this issue.

Here the term “sides” has been used instead of “pages,” as manuscripts are normally counted by leaves (= folios). Each folio has two sides: recto (facing side) and verso (reverse side). These are customarily foliated by the librarian in pencil in the top right-hand corner when the manuscript is acquired, even where it contains original pagination. Note that “sheet” is a unit for counting paper as a physical object, whereas “folio” refers to the same physical object when forming part of a manuscript. (NB. “folio” may also be used to describe a particular paper size.)

Fascicles are described in Latin by the number of bifolios: unio (1), binio (2), ternio (3), quarternio (4), quinio (5), senio (6) and so on. They are also expressed in abbreviated form in Roman numerals, e.g., 6 IV + 2 + 3 II (6 quarternios, 2 single leaves, 3 binios). Note that the German term Faszikel does not have the same meaning: it refers to the entire gathering of fascicles created by the writer, the actual German equivalent of fascicle being Lage.

When Bach made a bifolio, he would have usually folded an uncut sheet once, a binio thus requiring two separate sheets. (Exceptions are limited to smaller bound oblong volumes, such as the Orgelbüchlein and the family anthologies discussed below.) Other composers, research has found, would have often cut one sheet and folded it to make two bifolios or a binio fascicle. This is particularly important for scholars studying the watermarks in Mozart’s scores; see Alan Tyson, Mozart: studies of the autograph scores (Cambridge: Harvard University Press, 1987), 5.

Marshall, Compositional Process, 1: 61–63. For a detailed observation of Bach’s compositional process, see Robert Marshall’s commentary accompanying the facsimile published jointly by Zentral-Antiquariat der DDR in Leipzig and Hänssler in Neuhausen-Stuttgart in 1983, reprinted in Robert Marshall, The Music of Johann Sebastian Bach. The Sources, The Style, The Significance (New York: Schirmer, 1989), 131–42.

See note 44.

NBA KB II/1a (2005): 16.

The present brown leather binding dates, according to Dürr, from the mid-nineteenth century; it probably replaced Bach’s more modest original binding, which may have consisted of paper boards, as is the case with Altnickol’s copy of WTC II. NBA KB V/6.1 (1989): 20.

NBA KB V/6.2 (1996): 81.

This manuscript was reported bound (“eingebunden”) when it was listed in CPEB NV, 66; see Peter Wollny, “Zur Überlieferung der Instrumentalwerke Johann Sebastian Bachs: Der Quellenbesitz Carl Philipp Emanuel Bachs,” BJ 82 (1996): 7–21, esp. 13–14.

Many of Bach’s original manuscripts were furnished with a wrapper (Umschlag), which kept together the score and a set of parts. The first page of the wrapper would have been made into a title page for the music. Later entries by subsequent owners, such as signatures, stamps, and auction lot numbers, which are very useful in researching reception history, may also have been added on the wrapper.

Auflagebogen is a bifolio manuscript made for performance purposes that is supposed to be used in an opened state (Pultanlage). It requires only one turn of the entire leaf (i.e., 2v+1r; 1v+2r), rather than two page turns necessary with a normal bifolio manuscript (1r, 1v+2r, 2v). Auflagebögen cannot be read in the correct musical sequence if they are bound at the spine. The so-called “London autograph” of WTC II is a typical example of unbound Auflagebögen, and P 416/4 is its copy.

The only known exception is the score of BWV 195 (D-B, Mus. ms. Bach P 65), which uses printed manuscript sheets. Kobayashi speculates that they were provided by the printer who prepared the publication of The Art of Fugue; see Yoshitake Kobayashi, Bach to no Taiwa: Bach Kenkyu no Saizensen / [Conversation with Bach: the Frontline of Bach Research] (Tokyo: Shogakukan, 2002), 147.

For a detailed discussion of Bach’s stave-ruling practices, see Marshall, Compositional Process, 44–7.

Taking the case of the Fifth Brandenburg Concerto, it would seem that Bach’s initial plan was to copy the first movement (219 bars) from the model (D-B, Mus. ms. Bach St 130) into the space of 24 pages (ff. 58r–69v) that were ruled with eight staves—six narrower staves for melody instruments and two wider staves for solo harpsichord below them. He needed to write roughly 9 bars of music per page (24 × = 216), but having copied the first five pages, he found it impossible to maintain this due to the extensive use of demisemiquaver (thirty-second note) runs in the harpsichord part. He was forced to revise his original one-system layout for the harpsichord cadenza. Instead of leaving blank staves for the tacet instruments, he reconfigured the staves on three pages (ff. 67r–68r) to a four-system layout, three narrow staves and a wide one. The remaining three pages (ff. 68v–69v) were occupied by the second movement, where the staves were configured in two systems (four plus four) of uneven appearance. This theory of Bach’s copying process emerged from a private discussion with Nobuaki Ebata, to whom I am very grateful.

See Marshall, Compositional Process, 47–61.

Dok 3: 3 (No. 631).

Christoph Wolff, “The Last Fugue: Unfinished?” Current Musicology 19 (1975): 71–77; also in Christoph Wolff, Bach. Essays on His Life and Music (Cambridge: Harvard University Press, 1991), 259–264 and 423.

This conclusion can be drawn from a study of the “London autograph” of WTC II. See Chapter 2 of Yo Tomita, “J. S. Bach’s Well-Tempered Clavier, Book II: A Study of its Aim, Historical Significance and Compiling Process,” (PhD diss., University of Leeds, 1990), and Yo Tomita and Richard Rastall, The Genesis and Early History of Bach’s Well-tempered Clavier, Book II: A Composer and his Editions, c.1738–1850 (Aldershot: Ashgate, forthcoming). It is quite possible that Bach replaced rastra every one or two months during the period from 1723 to 1725, when he was preparing weekly cantata performances, which generally featured newly composed works.

Christoph Wolff, “Die Rastrierungen in den Originalhandschriften Joh. Seb. Bachs und ihre Bedeutung für die diplomatische Quellenkritik,” Festschrift für Friedrich Smend zum 70. Geburtstag dargebracht von Freunden und Schülern (Berlin: Merseburger, 1963), 80–92. Although Emery criticized Wolff’s theory for the uncertainties surrounding the stave measuring (see ML 45 (1964): 167–170), the value and validity of rastorology undoubtedly remain. It was unfortunate that at the time of Wolff’s proposal, the only comprehensive attempt by Hans Otto Hiekel on Bach’s original manuscripts, Katalog der Rastrierungen in den Originalhandschriften Johann Sebastian Bachs (Göttingen, unpublished typescript, 1965) did not produce significant results, mainly due to the imprecise method of measurement. It should, however, be possible to produce a new and more reliable catalogue of Bach’s rastra using the latest information technology equipment, as discussed later in this chapter.

For example, D-B, Mus. ms. Bach P 25, f. 37v, lower system (BWV 244/29, at b. 69). NBA KB II/5 (1974): 24.

Kobayashi, Bach to no Taiwa, 149.

Masahiro Niitsuma, Lambert Schomaker, Jean-Paul van Oosten and Yo Tomita, “Writer Identification in Old Music Manuscripts Using Contour-Hinge Feature and Dimensionality Reduction with an Autoencoder,” Richard Wilson, et al., eds., Computer Analysis of Images and Patterns: 15th International Conference, CAIP 2013, York, UK, August 27–29, 2013, Proceedings (Berlin and Heidelberg: Springer, 2013), 2: 555–62.

Bruce Kusko, “Proton Milloprobe Analysis of the Hand-Penned Annotations in Bach’s Calov Bible,” Howard H. Cox, ed., The Calov Bible of J. S. Bach (Ann Arbor: UMI Research Press, 1985), 31–106, esp. 42.

Uwe Wolf, Oliver Hahn and Timo Wolff, “Wer schrieb was? Röntgenfluoreszenzanalyse am Autograph von J. S. Bachs Messe in h-Moll BWV 232,” BJ 95 (2009): 117–34; Uwe Wolf, “Many Problems, Various Solutions: Editing Bach’s B-minor Mass,” Yo Tomita, Robin A. Leaver and Jan Smaczny, eds., Exploring Bach’s B-minor Mass (Cambridge: Cambridge University Press, 2013), 165–85, esp. 178–184.

For example, the review of Rainer Kaiser’s doctoral thesis “Tintenanalytische Untersuchungen an Bachschen Originalhandschriften mit Hilfe reflexionsspektroskopischer Methoden. Ein Beitrag zur archäometrischen Bach-Forschung” (Berlin: Technische Universität, 1987) in Yoshitake Kobayashi’s monograph Bach: Densho no Nazo o ou / [Bach: Pursuing the Mystery of Transmission] (Tokyo: Shunju sha, 1995), 21, and my own experiments on Bach’s autograph of BWV 872/2 as reported in “Analysing Bach’s ink Through a Glass Darkly,” MT 139 (Winter 1998): 37–42, esp. 41–2.

Wolff, “The last fugue unfinished?” 71–7.

Kobayashi, Bach to no Taiwa, 154. A sketch of a fugue in pencil exists; see Marshall, Compositional Process, 1: 32; 2: No. 150.

Spitta’s description of Anna’s handwriting is given in Spitta 1: 754–755; Spitta ET 2: 147–148. It should be noted, however, that some of his attributions to her appear to be incorrect: for example, the portion of Brockes Passion (D-B, ms. 9002/10) was, according to Dürr Chr 2, 119 and Kobayashi, 33 and 62, copied by principal copyist (Hauptkopist) H (= Johann Nathanael Bammler, identified by Wollny, “Neue Bach-Funde,” BJ 83 (1997), 36ff; see also NBA IX/3, 173ff.), while the score of BWV 165 (D-B, Am. B. 105) was, according to Dürr Chr 2, 71 and 150, copied by Anon. Id (= Johann Christian Köpping, identified by Schulze, “Beiträge zur Bach-Quellenforschung,” Carl Dahlhaus, Reiner Kluge, Ernst Hermann Meyer and Walter Wiora, eds. Gesellschaft für Musikforschung. Bericht über den Internationalen Musikwissenschaftlichen Kongress Leipzig 1966 (Kassel: Bärenreiter, 1970), 269–275; see also Dürr Chr 2, 173 n.51 and NBA IX/3, 55f.).

Identified by Hans-Joachim Schulze, “Johann Sebastian Bach und Christian Gottlob Meißner,” BJ 54 (1968): 80–88; also see Schulze, 101–110 and Dürr Chr 2, 163, n.4. In the 1950s Meißner was known as “Anon. 1” (Dadelsen, TBSt 1) and Hauptkopist B (Dürr Chr).

“Freilich zeigen seine Schriftzüge zunächst noch keine sonderliche Verwandtschaft mit denen Bachs. Erst im Laufe der Jahre beginnt sich ein Zeichen nach dem andern nach dem Leitbild der Bachschen Handschrift umzuformen; und wir wären schwerlich bereit, in den Ausgangs- und Endformen denselben Schreiber wiederzuerkennen, ließe sich nicht der Übergang der einzelnen Formen laufend belegen. Dabei vollziehen sich die Veränderungen im Gegensatz zu denen in der Schrift des Hauptkopisten A meist kontinuierlich, weniger sprunghaft; doch sind sie so zahlreich, daß sich besonders für die erste Zeit seines Auftretens eine Reihe charakteristischer Entwicklungsstadien aufzeigen läßt.” Dürr Chr 2, 26.

Dadelsen TBSt 1, esp. 32–33; Yo Tomita, “Anna Magdalena as Bach’s Copyist,” 59–76, esp. 67; NBA IX/3, 20–21.

For the summary of this development, see Dadelsen TBSt 1: 10–11, and NBA IX/3 (2007): xix–xxi.

Walter Emery, “The London Autograph of ‘The Forty-Eight,”’ ML 34 (1953): 106–123.

Dadelsen, TBSt 1 and TBSt 4/5.

Dadelsen, TBSt 1. This methodology was subsequently significantly refined by Kobayashi; see NBA IX/3 (2007): xxi–xxiii.

Young age until ca. 1713, the second half of Weimar years 1714–16, Cöthen 1717–23, Leipzig (1723 to January 1724; 1724–34; 1735–ca. 1744/46; the final years).

See note 3; see the summary in NBA IX/2 (1989).

See, for example, the added sharp in b. 78 of “Et in unum Dominum” in Bach’s autograph of the B-minor Mass (D-B, Mus. ms. Bach P 180) as discussed by Joshua Rifkin in “Blinding Us with Science? Man, Machine and the Mass in B Minor,” ECM 8 (2011): 77–91, esp. 78.

Dürr St; see also Dürr St 2.

Dürr Chr 2.

Hauptkopist A (Anon. 3) and Hauptkopist B (Anon. 1) were later identified as Johann Andreas Kuhnau and Christian Gottlob Meißner respectively by Schulze. See Dürr Chr 2, 163; see also NBA IX/3 (2007): xx n.9.

The result was essentially the same as that of Dadelsen. See Georg von Dadelsen, “Bachforschung in Tübingen,” Arnold Feil and Thomas Kohlhase, eds., Über Bach und anderes: Aufsätze und Vorträge 1957–1982 (Laaber: Laaber, 1983), 159–167, esp. 163.

Kobayashi. See also his two catalogues, NBA IX/2 and IX/3 (with Kirsten Beißwenger) that, together with NBA IX/1, manifest the result of research on Bach’s original manuscripts carried out in the second half of the twentieth century. See “Bach’s copyists” in this chapter.

Klaus Hofmann, “Bach oder nicht Bach? Die Neue Bach-Ausgabe und das Echtheitsproblem. Mit einem Beitrag von Yoshitake Kobayashi über ‘Diplomatische Mittel der Echtheitskritik,’” Hanspeter Bennwitz, ed., Opera incerta: Echtheitsfragen als Problem musikwissenschaftlicher Gesamtausgaben. Kolloquium Mainz 1988, Bericht. Akademie der Wissenschaften und der Literatur, Mainz. Im Auftr. des Ausschusses für Musikwissenschaftliche Editionen der Konferenz der Akademien der Wissenschaften in der Bundesrepublik Deutschland (Stuttgart: Steiner, 1991), 9–48.

Alfred Dürr, “Zur Entstehungsgeschichte des 5. Brandenburgischen Konzerts,” BJ 61 (1975): 63–69, esp. 64–65.

NBA KB VII/2 (1956): 103.

On the evidence of the similarities between their handwritings, Wollny suggests that Farlau studied with Altnickol in Naumburg before the former had commenced his university studies at Jena University in 1756. Peter Wollny, “Tennstädt, Leipzig, Naumburg, Halle: Neuerkenntnisse zur Bach-Überlieferung in Mitteldeutschland,” BJ 88 (2002): 29–60, esp. 43–44.

Through handwriting analysis, Kobayashi dated the first thirteen ca. 1739–42, and the remaining two ca. 1746–7. Kobayashi, 45 and 56.

Peter Wollny, “Einführung” and “Introduction,” Peter Wollny, ed., Die Achtzehn Grossen Orgelchoräle BWV 651–668 und Canonische Veränderungen über ‘Vom Himmel Hoch’ BWV 769. Faksimile der Originalhandschrift (Laaber: Laaber, 1999), ix (German) and xvii (English).

That is, after the Canonic Variations (BWV 769a).

This copy was split into four portions after Bach’s death, presumably while in the possession of W. F. Bach: (1) D-B, Mus. ms. Bach P 416, Faszikel 4; (2) Newberry Library, Chicago, Case MS 6A 72; (3) GB-Lbl Add MS 38068; and (4) the portion once housed in D-Dl lost during WWII; see also Kobayashi, 29–30, and NBA IX/3 (2007): 171.

These can be identified by studying the ink. Some of them were later scratched out; for details, see Yo Tomita, J. S. Bach’s ‘Das Wohltemperierte Clavier II’: A Critical Commentary, vol. I: Autograph Manuscripts (Leeds: Household World, 1993).

Kobayashi, Bach to no Taiwa, 260, 344–345; Yoshitake Kobayashi, “Exkurs über den Schreiber Anonymus Vr bzw. Anonymus 12,” Siegbert Rampe, ed., Bach: Das Wohltemperierte Klavier I: Tradition, Entstehung, Funktion, Analyse. Ulrich Siegele zum 70. Geburtstag (Munich: Katzbichler, 2002), 14–16.

The latest version of BWV 195 (D-B, Mus. ms. Bach St 12) is dated by Kobayashi, 61–3, “between August 1748 and October 1749”; the fourth version of BWV 245 (D-B, Mus. ms. Bach St 111) was prepared for the performance on 4 April 1749.

The score of Wq. 215 (D-B, Am. B. 170), copied on Berlin paper (NB. Wq 215 was composed in August 1749); Wq 65/25, 65/27 (on Berlin paper), 65/30 (with C. P. E. Bach’s corrections; the work was composed in 1758). These were found in D-B, Mus. ms. Bach P 789 and P 776. See Christine Blanken, “Zur Werk- und Überlieferungsgeschichte des Magnificat Wq 215 von Carl Philipp Emanuel Bach,” BJ 92 (2006): 229–71, esp. 236f.; also Kobayashi, 30.

Dok 1 (No. 50).

Dok 2 (No. 587).

Dok 3 (No. 640). Their second child, named Augusta Magdalena, was born and baptized in Naumburg with Anna Magdalena and Johann Gottfried Müthel among the godparents.

Christoph Wolff, “The Deathbed Chorale: Exposing a Myth,” Bach: Essays on His Life and Music, 282–294.

Yoshitake Kobayashi, “Zur Teilung des Bachschen Erbes,” Acht kleine Präludien und Studien über BACH. [Festschrift für] Georg von Dadelsen zum 70. Geburtstag am 17. November 1988 (Wiesbaden: Breitkopf & Härtel, 1992), 67–75. See also Peter Wollny, “Überlegungen zur Bach-Überlieferung in Naumburg,” BJ 86 (2000): 87–100.

Forkel names Altnickol as the person to whom Bach dictated this chorale. NBR, p. 466; Dok 7, p. 66 n.143.

Kobayashi, Bach to no Taiwa, 260.

Martin W. B. Jarvis, “Did Johann Sebastian Bach Write the Six Cello Suites?” (PhD diss., Faculty of Law Business & Arts, Charles Darwin University, 2007); see also Jarvis, “The Application of Forensic Document Examination Techniques to the Writings of J. S. Bach and A. M. Bach,” UB 3 (2008): 87–92. Here the point is to draw attention to the possibilities presented by a new methodology, not the erroneous conclusion drawn from it; see Ruth Tatlow, “A Missed Opportunity: Reflections on Written by Mrs. Bach,” UB 10 (2015): 141–157.

Masahiro Niitsuma and Yo Tomita, “Classifying Bach’s Handwritten C-clefs,” Anssi Klapuri and Colby Leider, eds., Proceedings of the 12th International Society for Music Information Retrieval Conference, ISMIR 2011, Miami, Florida, USA, October 24–28, 2011 (Miami: University of Miami, 2011), 417–21.

NBA KB I/35 (1964, Alfred Dürr, ed.): 40–41, and NBA KB I/14 (1963, Alfred Dürr and Arthur Mendel, eds.): 106. One of the few exceptions is the harpsichord part of BWV 1050 in Bach’s handwriting of ca.1721 (D-B, Mus. ms. Bach St 130), where the pitch of f#1 indicated by the key signature is cancelled by a flat in bb. 198–9.

See, for example, BWV 244/29, bb. 51–3: f. 36v in the autograph.

See BWV 244/59, b. 2 in the alto: f. 67v in the autograph. NBA KB II/5 (1974): 209.

Dok 3: 77 (No. 663). For a comprehensive study on Ringk’s copying of Bach’s works, see Dietrich Killian, NBA KB IV/5+6 (1978): 198–205; see also Russell Stinson, The Bach Manuscripts of Johann Peter Kellner and His Circle. A Case Study in Reception History (Durham: Duke University Press, 1989.

See the table listing the instances of flats used in place of natural signs as well as uses of the old semiquaver form and in 37 manuscripts copied by Ringk in Killian, NBA KB IV/5+6 (1978): 204: it shows that Ringk’s copy of BWV 565 was written before ca. 1740. See also Rolf Dietrich Claus, Zur Echtheit von Toccata und Fuge d-moll BWV 565, 2nd ed. (Cologne: Dohr, 1998), in which Claus argues against this view from the standpoint of style criticism.

Strictly speaking, the copyist who made the entire set of parts (without duplicates) from a score is called the Hauptschreiber. Hauptkopist is a specific term proposed by Dürr to refer to the scribe whom Bach employed to act as Hauptschreiber. Bach scholars have hence taken a stance to recognise those main scribes as Hauptkopisten when there is evidence of their making of multiple sets of parts.

Dürr Chr 2, esp. 145f.

Dürr Chr 2. Nearly all the Hauptkopisten have been identified, the most recent being Hauptkopist H = Johann Nathanael Bammler. Wollny, “Neue Bach-Funde,” 36–50. On the contrary, the identification of assistant copyists (Nebenkopisten) proved to be more difficult, because of a relatively fewer number of known specimens of their handwriting.

Yoshitake Kobayashi and Kirsten Beißwenger, Die Kopisten Johann Sebastian Bachs. Katalog und Dokumentation, 2 vols. (Kassel: Bärenreiter, 2007). Concordances of copyists with various anonymous codes given by earlier scholars, including Paul Kast and Eva Renate Blechschmidt, are given on pp. 216–218. In addition, Maul suggests that more names might be identified by a careful study of a recently uncovered notebook that belonged to a pupil of St Thomas.’ Michael Maul, “‘welche ieder Zeit aus den 8 besten Subjectis bestehen muß’ – Die erste ‘Cantorey’ der Thomasschule - Organisation, Aufgaben, Fragen,” BJ 99 (2013): 11–77, esp. 15–19.

The Bach Digital’s online source catalogue ( does contain a certain amount of such information and can be used for this purpose.

For example, Schulze, 46f.; Hans-Joachim Schulze, “Johann Christoph Bach (1671–1721), ‘Organist und Schul Collega in Ohrdruf’, Johann Sebastian Bachs erster Lehrer”, BJ 71: 55–81; Robert Hill, “The Möller Manuscript and the Andreas Bach Book: Two Keyboard Anthologies from the Circle of the Young Johann Sebastian Bach,”’ (PhD diss., Harvard University, 1987), 2 vols.

For example, Hermann Zietz, Quellenkritische Untersuchungen an den Bach-Handschriften P 801, P 802 und P 803 aus dem ‘Krebs’schen Nachlass’ unter Berücksichtigung den Choralbearbeitung des jungen J. S. Bach (Hamburg: Wagner, 1969).

For example, Hans Löffler, “Johann Tobias Krebs und Matthias Sojka, zwei Schüler Johann Sebastian Bachs,” BJ 37 (1940–48): 136–148.

For example, Hans-Joachim Schulze, “‘Das Stück in Goldpapier’: Ermittlungen zu einigen Bach-Abschriften des frühen 18. Jahrhunderts,” BJ 64 (1978): 19–42, and Schulze, 59ff.

For example, Alfred Dürr, “Heinrich Nicolaus Gerber als Schüler Bachs,” BJ 64 (1978): 7–18.

Kayser has long been known as Anonymous 5 (= Anon. Ih), whose handwriting was very similar to Bach’s. He was identified by Talle. Andrew Talle, “Nürnberg, Darmstadt, Köthen: Neuerkenntnisse zur Bach-Überlieferung in der ersten Hälfte des 18. Jahrhunderts,” BJ 89 (2003): 143–172 esp. 155f.

Dürr Chr 2, 44–65.

For example, Ulrich Leisinger, “Johann Christian Kittel und die Anfänge der sogenannten späteren thüringischen Bach-Überlieferung,” Rainer Kaiser, ed., Bach und seine mitteldeutschen Zeitgenossen. Bericht über das internationale musikwissenschaftliche Kolloquium, Erfurt und Arnstadt 13. bis 16. Januar 2000 (Eisenach: Wagner, 2001), 235–251.

See, for example, Kobayashi Yoshitake, “Franz Hauser und seine Bach-Handschriftensammlung,” (PhD diss. University of Göttingen, 1973), 109–112.

Dürr, “Zur Chronologie der Handschrift Johann Christoph Altnickols und Johann Friedrich Agricolas,” and NBA KB II/5 (1974): 62f.

Wollny, “Tennstädt, Leipzig, Naumburg, Halle,” 36–47.

Boston Public Library M.200.12 (2); NBA KB V/2 (1981): 39–47.

Bach-Archiv, Leipzig, Go. S. 19; Yo Tomita, “Bach and His Early Drafts: Some Observations on Little Known Early Versions of Well-Tempered Clavier II and the Goldberg Variations from the Manfred Gorke Collection,” BACH 30/2 (1999): 49–72.

This includes the internal logic of the composition itself, which the composer felt was still lacking, and any external circumstances, such as the setting of a piece for a new context, which may necessitate transcribing or transposing it for a different instrument or expanding it to fit into a new work, as is the case with many movements of WTC II.

The best-known examples are the early versions of Inventions in C minor (BWV 773) and G minor (BWV 797) found in the Clavierbüchlein for Wilhelm Friedemann, in which we find evidence of Bach adjusting the melodic shape of the theme to ensure dramatic cohesiveness.

See the variants of the A-minor prelude in WTC II (BWV 889/1), b. 30. The initial reading found in his autograph (GB-Lbl, Add. MS 35021) gives a passage of weaker contrapuntal logic, which Bach adjusted in his later (lost) copy, as can be observed in Altnickol’s 1744 copy (D-B, Mus. ms. Bach P 430), among others.

See the textural clarification Bach makes in the Prelude in F minor in WTC II (BWV 881/1), bb. 38–9; compare the Prelude in F-sharp minor in WTC II (BWV 883/1), b. 20.

See, for example, how BWV 831 was gradually clarified notationally. Walter Emery, “An Introduction to the Textual History of Bach’s Clavierübung, Part II,” MT 92 (May 1951): 205–209; and 92 (June 1951): 260–262.

See, for example, the slur provided for the demisemiquaver (thirty-second note) flourish in the bass in b. 21 of BWV 870/1 in Altnickol’s 1744 copy: there is no slur in the corresponding passage in the earlier part of the prelude in b. 6. Logically, the placement should be reversed. A possible explanation of such an anomaly is that additions were made during a lesson.

Joshua Rifkin, “Bach’s Chorus: A preliminary Report,” MT 123 (November 1982): 747–754; see also Andrew Parrott, The Essential Bach Choir (Woodbridge: Boydell, 2000).

See, for example, Uwe Wolf, “Von der Hofkapelle zur Stadtkantorei: Beobachtungen an den Aufführungsmaterialien zu Bachs ersten Leipziger Kantatenaufführungen,” BJ 88 (2002): 181–191, esp. 187–191; for a fair summary of the ensuing debate, see Robin A. Leaver, “Performing Bach: One or Many?” Choral Scholar 1/1 (Spring 2009): 6–15.

One such example that came to light with recent source recoveries are the ripieno parts of Cantata 23 that once completed the original set. The parts were separated after the Berlin Singakademie obtained the manuscript. See Christoph Wolff, “Zurück in Berlin: Das Notenarchiv der Sing-Akademie. Bericht über eine erste Bestandsaufnahme,” BJ 88 (2002): 165–169.

This includes research on Bach’s changing attitudes to performances in response to varying and evolving circumstances, iconographical evidence found in the engravings in the contemporary sources, and documentary evidence referring to the size of choir and how the choir held the parts sharing or not sharing with fellow singers. One must be aware that the weight of support gained from this type of evidence is still very small and requires further systematic study.

For example, John Butt, Bach Interpretation. Articulation Marks in Primary Sources of J. S. Bach (Cambridge: Cambridge University Press, 1990).

For example, Masaaki Suzuki has written “Seisaku Noto /制作ノート” (Production Notes) in the concert program booklets of Bach Collegium Japan since 1995, in which he often discusses the original parts in the context of his own interpretation.

Frederick Neumann, “Some Performance Problems of Bach’s Unaccompanied Violin and Cello Works,” Mary Ann Parker, ed., Eighteenth-Century Music in Theory and Practice: Essays in Honor of Alfred Mann (Stuyvesant: Pendragon Press, 1994), 19–46, esp. 37. For a discussion from a source-critical perspective on wider editorial issues, see Georg von Dadelsen, ‘Die Crux der Nebensache – Editorische und praktische Bemerkungen zu Bachs Artikulation’ BJ 64 (1978): 95–112.

See “Commentary” in Bettina Schwemer and Douglas Woodfull-Harris, eds., 6 Suites a Violoncello Solo senza Basso. BWV 1007–1012. / J. S. Bach (Kassel: Bärenreiter, 2000), 6–7.

See NBA II/5 KB, 129–130 (Dürr); see also Dürr “De vita cum imperfectis.”

Kobayashi, Bach to no Taiwa, 224–227.

Schulenberg’s explanation appears to be the most likely. He interprets the dots added in the B-flat minor fugues “to mean accentuation and separation, rather than staccato in the modern sense,” and to “reinforce the subject’s declamatory quality.” David Schulenberg, The Keyboard Music of J. S. Bach, 2nd ed. (New York: Routledge, 2006), 271.

For further discussion, see Yo Tomita, “Reading Soul from Manuscripts: Some Observations on Performance Issues in J. S. Bach’s Habits of Writing His Music,” Thomas Donahue, ed., Essays in Honor of Christopher Hogwood. The Maestro’s Direction (Lanham: Scarecrow, 2011), 13–40.

Modern critical editions, including NBA, tend to normalize inconsistencies in beaming, treating it as something that requires editorial arbitration, even though their editorial guidelines advise editors to retain Bach’s own grouping of notes under beams as it may convey articulation. See the section “Balkensetzung” in “Editionsrichtlinien. Johann Sebastian Bach. Neue Ausgabe sämtlicher Werke,” Georg von Dadelsen, ed., Editionsrichtlinien musikalischer Denkmäler und Gesamtausgaben (Kassel: Bärenreiter, 1967), 61–80, here 66; revised and updated version in Frauke Heinze and Uwe Wolf, Gesamtregister, NBA IX/4 (2010): 319–335 at 325: “Da die Bachsche Notengruppierung unter Umständen für die Artikulation von Bedeutung ist, ist sie nach Möglichkeit beizubehalten.” Indeed, Bach’s notation of quaver beaming is often inconsistent, and the problem is felt even more acutely when dealing with copies. Although a reliable way of extracting such information is yet to be devised, the possibility of rediscovering Bach’s intentions should not be ignored. Future critical editions must address this issue seriously. For further discussion on this issue, see Tomita, “Reading Soul from Manuscripts” and “Deciphering J. S. Bach’s Performance Hints Hidden in J. S. Bach’s Quaver Beams, ” EM 44 (2016): 89–104.

See Chapter 6, “Bach as Teacher” in David Ledbetter, Bach’s Well-tempered Clavier: The 48 Preludes and Fugues (New Haven: Yale University Press, 2002), 126–140.

CPEB NV 66–69. Note that the third item was listed at the very end of Bach’s vocal works (“Singstücke”) on p. 81. See also Dok 3: 491–501 (No. 957).

Most likely D-B, N. Mus. ms. 10490, containing BWV 870a, 899, 900, 901 and 902, a collection first publicly discussed by Klaus Hofmann “‘Fünf Präludien und fünf Fugen’: Über ein unbeachtetes Sammelwerk Johann Sebastian Bachs,” Winfried Hoffmann and Armin Schneiderheinze, eds., Bach-Händel-Schütz-Ehrung 1985 der Deutschen Demokratischen Republik. Bericht über die Wissenschaftliche Konferenz zum V. Internationalen Bachfest der DDR in Verbindung mit dem 60. Bachfest der Neuen Bachgesellschaft. Leipzig, 25. bis 27. März 1985 (Leipzig: VEB Deutscher Verlag für Musik, 1988), 227–235; see also James Brokaw, “Recent Research on the Sources and Genesis of Bach’s Well-tempered Clavier, Book II,” BACH 16/3 (July 1985): 17–35. The pieces in question were published together with WTC II in NBA V/6.2 (1995).

Most likely D-B, Mus. ms. Bach P 563, containing BWV 844a, 933, 872a/2 (in C), 901/1, Wq 111.

Most likely D-B, Mus. ms. Bach P 226, containing BWV 872a/1 (in C), 875 and Fk 26–28.

See, for example, Mark Lindley “Early fingering: Some Editing Problems and Some New Readings for J.S. Bach and John Bull,” EM 17 (1989): 60–69, esp. 65; Peter Le Huray, “Bach’s C major Prelude BWV 870 and 870a,” in Peter Le Huray, ed., Authenticity in Performance: Eighteenth-Century Case Studies (Cambridge: Cambridge University Press, 1990), 5–23, esp. 10f.

See Alfred Dürr, “Ein Dokument aus dem Unterricht Bachs?” Musiktheorie 1/2 (1986): 163–170.

For further discussion, see “Transmission and Reception History” in this chapter.

Original musical text does not necessarily imply a single version. Various issues may arise, ranging from differences in versions to discrepancies in readings between the score and performance parts.

For example, Wolf, Hahn and Wolff, “Wer schrieb was?” 117–134.

A revised reading can be difficult to decipher accurately. For example, the natural Bach wrote on beat 2 of bar 10 of the Sarabande in Partita 2 in D minor (BWV 1004/3) has frequently been misinterpreted by scholars as a flat. See Neumann, ‘Some Performance Problems of Bach’s Unaccompanied Violin and Cello Works,” 22–23, and Martin Jarvis, “Is Anna Magdalena’s Manuscript of the Violin Sonatas & Partitas a Copy of the ‘1720 Autograph’ of the Unaccompanied Violin Sonatas & Partitas of J. S. Bach?” Stringendo 29/2 (2007): 41–42.

See Dadelsen, “Editionsrichtlinien,” 63, and Heinze and Wolf, Gesamtregister, 333.

However, in situations where multiple sources exist for the same version of a piece, it is essential to assess every variant in order to determine which readings originated from the composer and which did not, or which reading the composer decided to replace with a new one. Choosing a single source in such a situation and producing its text faithfully to avoid conflation of text represents a mere transcription (with no editing involved), and therefore cannot be regarded as a critical edition. For this reason, I would remonstrate against the editorial principles adopted in Peters’ The Complete Chopin: A New Critical Edition.

Perhaps the best-known example is the 1720 autograph of Violin Solos. In b. 3 (the beginning of the second system) of the opening G-minor Adagio (BWV 1001/1), Bach omits the flat for e e 1. Numerous examples can also be found in the original manuscript of WTC II, hinting that earlier versions of some movements were written in modal key signatures.

Two specific instances are found in WTC II: Fugue in D major (BVW 874/2; autograph missing), b. 45 tenor (g – Bach, for a while, forgot to write the sharp as the passage modulated to A major); Fugue in G minor (BWV 885/2), b. 64 alto (a a 1 – Bach forgot to add the flat as the passage modulated to C minor).

Strictly speaking, the eighth volume of its supplement series Bach-Dokumente, “Materialien zur Quellen-Überlieferung bis 1850,” is yet to appear.

For further discussion on this issue, see Yo Tomita, ‘“Rotancho Misakyoku” to Bach Kenkyu: Sakuhin no Densho to Gakufu Henshushi kara miru Kenkyujo no Shomondai’ /《ロ短調ミサ曲》とバッハ研究: 作品の伝承と楽譜編集史からみる研究上の諸問題 [The B-minor Mass and Bach Research: Various research questions posed by the work’s transmission and the history of editing], Ongaku Kenkyujo Nenpo / Kunitachi College of Music Research Institute Bulletin 24 (March 2012): 145–182, esp. 164.

By carefully examining the stemma of secondary sources it is sometimes possible to discover the variants once contained in a lost original manuscript, which would otherwise be inaccessible. Charting the process of transmission can thus become a twoway street, at the same time moving ahead towards subsequent related sources and back towards the composer’s original text.

Tanja Kovačević and Yo Tomita, “Neue Quellen zu Johann Sebastian Bachs Violinsoli (BWV 1001–1006). Zur Rekonstruktion eines wichtigen Überlieferungszweigs,” BJ 95 (2009): 49–74.

Yo Tomita, “Breaking the Limits: Some Preliminary Considerations on Introducing an e-Science Model to Source Studies,” Ongakugaku to Globalization: Nihon Ongaku Gakkai Soritsu 50 Shunen Kinen Kokusai Taikai: Shizuoka 2002. Hokokusho = Musicology and globalization: proceedings of the International Congress in Shizuoka 2002 in celebration of the 50th anniversary of the Musicological Society of Japan (Tokyo: Nihon Ongaku Gakkai, 2004), 233–237; see also Tomita, “Challenging Virtuality,” 4–6.

Two of these canons (nos. 11 and 13) were known in slightly different versions, BWV 1077 and 1076 respectively.

For an account of the circumstances of its discovery, see Olivier Alain, “Un supplément inédit aux Variations Goldberg de J. S. Bach,” Revue de Musicologie 61 (1975): 244–245; see also Christoph Wolff, “Bach’s Handexemplar of the Goldberg Variations,” JAMS 29 (1976): 224–241. This copy is presently held by the Bibliothèque Nationale de France in Paris under the shelfmark: MS 17 669.

Carl Philipp Emanuel Bach: The Complete Works published by the Packard Humanities Institute in Los Altos, CA.

Michael Maul, “‘Alles mit Gott und nichts ohn’ ihn’ – Eine neu aufgefundene Aria von Johann Sebastian Bach,” BJ 91 (2005): 7–34.

Michael Maul and Peter Wollny, eds. Weimarer Orgeltabulatur. Die frühesten Notenhandschriften Johann Sebastian Bachs sowie Abschriften seines Schülers Johann Martin Schubart. Mit Werken von Dietrich Buxtehude, Johann Adam Reinken und Johann Pachelbel, Faksimile-Reihe Bachscher Werke und Schriftstücke. Neue Folge, III; Documenta Musicologica. Zweite Reihe: Handschriften-Faksimiles, 39 (Kassel: Bärenreiter, 2007).

A revised version is published in UB 10 (2015): 123–137.

This research should also include the eastern areas of formerly German lands (Prussia) that are parts of Poland and Russia today where Bach’s students worked after studying with Bach.

Thekla Kluttig, “Nur Briefe berühmter Komponisten? Archivgut von Leipziger Musikverlagen als Quelle für die Musikwissenschaften,” Mf 66/4 (2013): 391–407.

Christine Blanken, “Ein wieder zugänglich gemachter Bestand alter Musikalien der Bach-Familie im Verlagsarchiv Breitkopf & Härtel,” BJ 99 (2013): 79–128..

See the forthcoming article by Nobuaki Ebata in UB.

Search for more...
Back to top

Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.