fourth installment, which examined the two “AMB” notebooks. Toward the end of that post, writing about one of the two incomplete copies of the aria “Schlummert ein” (from BWV 82) in one of these notebooks, I wrote:
Several pages later the aria appears a second time, although this time AMB did not finish the copy. The vocal line breaks off midway through bar 60 (at the end of a page); the unfigured bassline breaks off after 28 bars. It seems likely to me that it was added in later, as it too breaks off at a page-turn: waiting for the ink to dry before turning the page, she was needed elsewhere and never completed the project. (Similarly, I wondered, are the five missing appoggiaturas in her first copy merely a sign of a practical notational issue? That is, might she have used a different pen-nib for the appoggiaturas, so that there was a reason to leave space and move on, coming back to fill them in later? I don’t know the Bach literature well enough to know if this has been explored, nor have I seen it discussed in other eighteenth-century sources.) [On the image on the left, the vertical blemish in the middle of my red circle where the appoggiatura ought to be does not seem to be an erasure—and there is no such blemish in the other four instances.]
|SOURCE: cropped scans of "Schlummert ein" b. 40 in AMB2 p. 108 (f. 55r) [with absent appoggiatura highlighted] and 113 (f. 59v) from Bach-Digital.
The example above is in the hand of Anna Magdelena Bach. In the past several months of exploring Bach sources (particularly in his own hand), this issue about the absence of small-note ornaments has recurred with such regularity that I find myself with a short catalogue of data points. I should stress that I’m thinking exclusively of the “small-note” ornaments (i.e., those in which a note is written smaller to indicate that it is ornamental) rather than what one might call the “squiggle” ornaments (i.e., those indicated by an arbitrary symbol—trill, mordent, turn, Schleifer, etc.) These latter show up regularly in composing scores, and thus (I am speculating) did not require a change of nib to notate. (And nota bene: Pen-nib was not quite accurate in my earlier post, as really it is quill-nibs that are at issue.)
Indeed, much of this post is really just speculation, as I do not have much to go on. I remember a moment in a graduate school seminar when I voiced some idea for which I had assembled a similar paucity of evidence, and the professor (rightly) shot it down with the line “That’s not a theory; that’s speculation.” The contempt with which he enunciated the word still sticks with me. For some reason, I enjoy reliving that moment in my memory—maybe because it was an important lesson I needed to learn.
I have also written in this blog about the dangers of amateurism in music scholarship—yet I will wallow in amateurism in this post. As I’ve said before, I am no Bach scholar; his music is an inevitable topic in a blog such as mine, as the amount of textual research to which his Nachlass has been subjected is truly staggering. There is just so much for me to write about. Like Everest, he’s there. But in this post I find myself writing speculatively, without any underpinnings in the literature.
|SOURCE: Bärenreiter promotional photo
I have made no general survey of the Bach works for which both a composing score and an authentic set of parts survive; that task is certainly beyond what can be done with a blog that is very much on the side of my obligations. But that the parts supply this sort of information when the scores may not is noted by Moritz Hauptmann, the original chairman of the Bach Gesellschaft edition, and the editor of this cantata in the second volume of the BG (1852):
|SOURCE: detail of "Vorwort" (p. xiii) to BG vol. 2, from scan available via the IMSLP.
[Third and fourth sentences; roughly:] “...Where in addition to the score the original parts were also available, these were conscientiously consulted. The parts are of importance not only for the appoggiatura markings and the figuring of the basso continuo, of which the score seldom has any; they also serve to verify unclearly-written notes and lyrics [in the score]....”A particularly striking example of this in BWV 20 is the sixth movement, the alto aria “O Mensch, errette deine Seele,” which is replete with small-note ornaments in both the BG and the NBA, both of which give proper editorial deference to the parts over the score in this respect. As a sample, observe (if you can make it out below) just the opening bars of the BG, the autograph, and the original Vln. I part:
|SOURCES: (top L) cropped screen-shot of BG Bd. 2, p. 314 from IMSLP; (top R) cropped scan of Bärenreiter 2017 facsimile of D-LEb Rara I,14; (bottom) cropped screenshot of original Vln. I part f. 2r. from Bach-Digital. [Note also that the trill in the score b. 3 is dutifully copied into the part; the trill in b. 2 is Bach’s added ornament (the identical mark as in the score) to the copied part, and thus not in the score.]
All of this brought to mind a startling difference between the early version of the St. Matthew Passion and the version universally familiar from Bach’s c. 1736 fair copy score. The sole extant source for the early version is a copy by J. C. Farlau made some ten or so years after Bach's fair copy “revised” version was prepared. Originally the NBA issued the early version only as a grey-scale facsimile of Farlau’s score (which at the time—1972—was attributed to Johann Christoph Altnickol), although it has subsequently been issued as a newly-edited volume in its own right. It has been recently recorded by the Academy of Ancient Music under Richard Egarr, who (on the promotional video for this release) notes the absence of some of these appoggiaturas:
“I find a lot of things sharper in focus and more dramatic in color, whereas in the later version things have been softened up with appoggiaturas and more rococo ornaments. So generally I find some things are a little bit more shocking, and a bit clearer.” (at 04:21)
On that video we hear what must be the most “shocking” example of this textual difference, which is the duet “So ist mein Jesus nun gefangen” (no. 27a, near the end of Part I). Here is the beginning as it appears in Bach’s c. 1736 manuscript, replete with "small notes" in the imitative woodwind lines:
|SOURCE: BWV 244 no. 27, bb. 1-5; detail (p. 55) of Bach's fair copy (D-B Mus. ms. Bach P 25) from Bach-Digital.
And the same passage as it appears in Farlau’s copy; nary an appoggiatura in sight:
|SOURCE: BWV 244b no. 27a, bb. 1-5; detail of copy by J. C. Farlau (D-B Am. B. 6) from Bach-Digital.
My speculative nib idea also suggests a new way at looking at (for example) the profusely ornamented chorale “Wenn wir in höchsten Nöten sein,” BWV 641 (in the Orgelbüchlein). This is the chorale which was later revised as the so-called “deathbed chorale” (“Vor deinen Thron tret’ ich,” BWV 668), stripped of the ornamentation, and with imitative interludes interpolated between the chorale phrases. (Christoph Wolff has dealt with this situation very lucidly.) But what strikes me as I look at the page is that the ornamented melody seems to have been written with two different sized quill-points—and am I imagining things to perceive a different tint to the ink? Here is the full page:
SOURCE: BWV 641, p. 115 of Bach's autograph of the Orgelbüchlein (D-B Mus. Ms. Bach P 283); scan from Bach-Digital
|And here is a detail of the third measure leading into the fourth. The downbeat of b. 3 seems to be a normal-sized note-head; all of the notes of beat two, and the first note of beat 3 and beat 4 similarly seem to be the default size, as well as all of the notes in the other voices. But the rest of the figuration seems significantly smaller:
|SOURCE: detail (bb. 3-4) of BWV 641 (D-B Mus. Ms. Bach P 283); scan from Bach-Digital
Turning to Peter Williams's excellent survey The Organ Music of J. S. Bach, I’m delighted to see that he has noticed the same thing:
“The coloraturas, unlike most of those in BWV 614 and 622, centre around turning phrases that lead to the next note of the cantus, which is placed where it would be even if there were no decoration. This is a particular technique that can be understood in two ways: these embellishments could be taken out in order to produce BWV 668, or they could be added in order to produce BWV 641, where they are written as smaller notes in P 283.” (p. 311; emphasis mine)
We might imagine (uh... speculate), then, that Bach initially wrote something like this:
only later to squeeze in the more florid version. His use of the smaller nib would then be merely a practical matter—not indicating “small-note ornamentation,” but simply cramming a lot of notes into an insufficient space. It is certainly hard to believe that he would have intended such a florid second beat of b. 4 when originally laying out the work in this manuscript, as he ends up not just in the margin but in the gutter of the binding of his “little book.”
Another imponderable question: do the squiggle ornaments which would then have been written mainly over quarter notes apply also to my putatively new florid line (i.e., resulting in the face-value reading of the page, as it is invariably played now)? Or does the filigree supersede the squiggles (or some of them), themselves mere remnants of a previous version?
One thought leads to another, and the suggestions here about BWV 641 strike me as support for those organists who would add ornamentation to another Orgelbüchlein chorale, BWV 639 “Ich ruf zu dir, Herr Jesu Christ,” in which the first half of the chorale is decorated with a few passing tones but the second half is left in bare quarter notes. I have always taken this to mean “you get the idea... go thou and do likewise.” The Orgelbüchlein was explicitly a teaching volume—and, like Bach's other pedagogical works, transmitted not in print but by students (and subsequently their students) copying it out. From the NBA critical report for this work and from perusing some of the scans available via Bach-Digital, I see that some of these extant copies (NBA sigla B2, L2 and M1 particularly) do indeed have a few added squiggle ornaments and passing tones in the second half—although, of course, the presence of such decoration on the page is not the prerequisite to the performance of a more decorated version.
And so I sit and speculate. And piece after piece comes before my eyes in which the similar small-note ornament discrepancies recur, although the conditions are never quite the same. It should be noted, for example, that solo music need not have the ornamental details worked out that ensemble music (particularly ripieno parts) would need. It is also worth keeping in mind that unlike our current rehearsal situation, where a pencil or other writing implement should always be handy, Bach's players may have had no writing implement at hand (surely not any requiring an inkwell!), and so it paid for ensemble parts to be as carefully prepared as time allowed.
I have been wondering about what evidence might refute this whole idea. Certainly an example intermingling squiggle and small-note ornaments in close proximity could be inconvenient for my proposal, although it might be hard to show that the two different sorts of notation were written in the source in the same sitting. (At this point in a draft of this post, I digressed with an example from BWV 873/1, the C# minor prelude from Book 2 of the Well-Tempered Clavier. I took it out because it was too dense for the flow of this already-too-dense post, but if you want to see it, here it is.)
I will close with an example I stumbled across while flipping through Walter Emery’s classic study, Bach’s Ornaments. As with most treatises on ornamentation, Emery’s subject is how an ornament should be interpreted in performance. That is not my subject, but I turned to Emery nonetheless because of his keen eye for the notation. I have quoted Emery on this blog before—and will surely do so again. In this book, he is scrupulously cautious about the sources he quotes, remarking sagely “one cannot deduce Bach's habits from ornaments that he did not write. As things stand, this means that unless a writer on ornamentation has made himself competent to edit every work he wishes to quote, he must take examples only from reproductions of autograph manuscripts, and from a few texts whose reliability can easily be tested. The Bach examples in this book have been chosen in this way; with a few exceptions, included for special reasons and expressly described as questionable, they are authentic beyond all reasonable doubt.” (p. 7)
Thus his Ex.157 caught my eye, in which small-note appoggiaturas surround a single hook (b. 11):
|SOURCE: cropped scan of Walter Emery, Bach's Ornaments (Novello, 1953), p. 77.
But Emery was cautious as ever:
“As the hooks have been translated into small notes by all but the most conscientious of editors, it is impossible to say, without access to large numbers of autograph manuscripts, whether Bach made any distinction between hooks and small notes. If the BG is to be trusted, he used them indifferently in the tenor aria of Cantata 67 [his Ex. 157].” (p. 76)
If the BG is to be trusted, he writes, and even before looking up the BG I went straight to Bach-Digital—a resource which gives precisely the “access to large numbers of autograph manuscripts” that Emery craved. There I found a scan of the original Violin I part, copied by J. A. Kuhnau, and what did I see?
|Source: BWV 67/ii, Violin I part, end of b. 1; cropped scan from Bach-Digital
Emery was right to be wary. In fact, BG is unreliable here: the small-note E appoggiatura is actually there—but I can’t help wondering if Kuhnau's “hooked” flags for the sixteenths might have caught the editor (Wilhelm Rust) out on this one instance—the hook from the f# flag somehow being turned into an appoggiatura hook on the d#. Otherwise it is hard to explain whence the hook in BG came. It does no damage in performance: the appoggiatura is authentic; it's just Rust's notation of it that is not. Even more curious, actually, is (for example) bar 5 of the same aria, where Rust converts the original small-notes all into hooks—rather than, as Emery would have expected, the other way around:
|SOURCE: BWV 67/ii b. 5-6; (top) BG vol. XVI, p. 228 (at IMSLP); (bottom) original Vl. I part (as above).
Rust actually does this switch inconsistently throughout the aria in both Violin and Oboe lines, and I can see no logic to it. The lesson I take from it, though, is that I don't have the time or the energy to plumb the depths of the small-nib question further. At least for now.
03 February 2019
Another example to keep in mind: BWV 125/ii: the original parts are preserved replete with small notes; but other sources transmit the presumed reading of the (now lost) score, lacking the graces. See NBA Ser. I Bd. 28.1, pp. 66-74 (cf. 46-54).