By Alexander Dean
In 2013 A-R published Dr. Andrew Levin’s edition of a viola concerto by a certain Joseph Schubert—court violist in the Dresden Hofkapelle in the late eighteenth and early nineteenth centuries (his name appears in a list of Dresden court musicians dating from 1800, under “Braccisten” [violists] as shown in figure 1).
| Figure 1. Churfürstlicher Sächsicher Hof- und Staats-Kalender auf das Jahr 1800 (Leipzig), page 63, showing the musicians employed in the Dresden Hofkapelle orchestra, excerpt. |
 |
The manuscript Dr. Levin presented to us was prepared from a set of parts he had uncovered in the Sächsische Landesbibliothek – Staats- und Universitätsbibliothek, Dresden. As usual for performing materials, there were some discrepancies among the parts. Now it is a common editorial process to sort out materials that stem from different performances, or from different times—such as, for instance, those that originate before a performance (the composer’s copy) and those dating from afterward (the parts as marked by the musicians, to reflect all the decisions made during rehearsal and performance). Also, performances might vary in size and instrumentation, depending on the venue, and the composer might make changes from one performance to the next.
The editor, in such cases, has to make some judgment calls. One may be tempted, at first, to strive to record everything at once in the modern edition, to faithfully conflate every conflicting reading into a mashup of critical notes, footnotes, addenda, marginalia, and as many various codes, sigils, and symbols that one’s ingenuity can summon. The result, of course, would be an artifact even more mysterious and obscure than the original sources themselves. A better course is to find a reasonable historical or scholarly basis on which to prioritize the source information, so that a single reading can be presented in modern notation. Having fixed on this reading, the other information in the source can then be presented in relation to it, as part of the critical apparatus.
Or at least that is the idea—reality often presents complications, as it did in the case of Dr. Levin’s Schubert edition. And this is not a bad thing—complicated reality is why historical editions are worthwhile, after all. In this case, the Dresden parts for Joseph Schubert’s concerto were already a bit of a concatenation: although they probably represent materials used for a single performance around 1800, they also betray fairly substantial changes made to the piece as originally envisioned. There are copies of the string parts, for instance, with alterations, including a segue between movements that does not appear in the “original,” unmarked set of parts; a third set of string parts, for a ripieno group, were copied with the segue material in place, rather than added at a later time. So, although all these copies appear in the same source, and most of them were probably used in the same performance, a progressive timeline seemed nevertheless to be discernable: the first set represents the original conception; the second set, copied for a performance according to this original conception, were altered to include expanded material; and then the ripieno parts were copied from those, with the expanded material included (figure 2 shows an example from each set).
| Figure 2. Examples of the three copyists’ work from the opening of the first movement, first violin. |
 |
And the presence of these ripieno parts suggests that these changes represent not only an expansion of material, but also of performing forces—a larger string section, which would be reduced while accompanying the solo viola. This is borne out by other clues: the wind and timpani parts are almost free of later edits and alteration, suggesting that they were added as part of the reconception of the piece, rather than altered to fit it; also, oboe cues in the “original” string parts are actually played by the clarinets, showing that the winds were rearranged as part of this expansion of the performing forces.
So what does this mean for the editor? How many “versions” of the piece are there in this case, and which should be preferred? And does the slippery term “original” present a help or a hindrance in conceptualizing these versions?
With our help, Dr. Levin found answers to these questions, and a way of demonstrating, in the modern edition, something of the complexity of the source material.
Bear in mind that, as already discussed, more complex is not always better: the goal is transparency, so that those aspects of the source that are judged to be important can show “through” the modern notation without obscuring it. In this case, the apparently chronologically ordered “versions” turned out to be something of a red herring; the scholarly impulse to seek out an “original” and then describe anything else in terms of that did not bear fruit. The Schubert parts were actually of a piece, surviving (probably) from a performance of the updated version, with only those sections that did not need an update left alone. There may well have been the conception, in the mind of the composer, of a smaller-scale concerto, without a ripieno section, with a slightly different grouping of wind instruments, and lacking certain short musical passages such as the segue between the second and third movements. But in the absence of concrete evidence that the piece was actually performed this way, and without the means to fully describe such a version, is it worthwhile to refer to it with the term “original”? This term, perhaps unwittingly, tends to prejudice us toward freighting the phantom version with a certain authenticity, one that might even be enhanced by its very elusiveness. The mysteriousness of the past, after all, is catnip to the historian. But in the end, the practical necessities of critical edition-making remind us of our duties: to confront history as it appears, which is in the present, in the materials that are here, now, in front of us. And those materials, in the case of the Schubert instrumental parts, while definitely layered, offer us only the end results of the piece’s history, not a model of its genesis. They manifest some of the birth pangs this concerto suffered as it came to take the form that can actually be described, and transcribed, on the basis of the surviving parts. They demonstrate, among other things, the composer’s response to the changing status of the ensemble and the genre in his time and place—a valuable discussion that rightly belongs in the prose treatment provided by Dr. Levin, augmented by descriptions of the source materials and critical notes, rather than embedded into the transcription.
By taking this more hard-headed approach, we were able to tackle a further, possibly more productive complication: alterations to the viola solo part representing yet another “layer”: bowings, fingerings, and embellishments probably added by the soloist, who may have been Schubert himself. These markings are superimposed on other, slightly different readings, presumably stemming from the time of composition. So while one might, using standard scholarly terminology, refer to these readings as “scribe A” versus “scribe B,” it is quite possible that they actually represent “Schubert as composer” versus “Schubert as performer.” (Dr. Levin’s full analysis of the different hands and versions at work is a bit more complicated than this, actually; see the introduction to the edition for the whole story.) And, as seen in figures 3 and 4, there we find catnip for the historically informed performer as well.
| Figure 3. Viola solo part showing performer’s alterations, first movement, measures 134–38. |
 |
| Figure 4. Viola solo part showing performer’s alterations, measures 253–78. |
 |
Some of the additions are explainable as adjustments complementing the aforementioned revisions to the orchestral parts; in fact, it is even possible that certain passages were added by the soloist and thus prompted the revisions to the other parts. But others are simply improvements and/or personalizations on the part of the soloist, neither more nor less “authentic” than the bowings, articulations, etc., set down when the part was first copied. Both layers transmit period-appropriate information on performance practice, and both are of potential interest to modern performers. But displaying both in the modern edition does present a certain challenge.
Of course, there are settled ways of differentiating various types of markings in a critical edition. But they are usually reserved for distinguishing the work of the editor from the readings in the source (placing brackets around elements added by the editor, for example)—keep in mind that such editorial markings do not necessarily represent a change to the source; they can serve to explicate something the source implies. For example, the Rondo movement, as usual for late-eighteenth- and early-nineteenth-century examples of the genre, especially in manuscript sources, gives articulations only for the first appearance of the theme, assuming that the many repetitions can be played the same way; other time-saving devices might include giving articulations, dynamics, etc., for one part, and letting them stand for other parts that follow in unison or are otherwise clearly connected musically. In a modern print, however, where the same style standards need to be used for the works of Schoenberg and Monteverdi alike, we have to be more punctilious: a lack of slurs means something, just as does their presence. And so we usually add the slurs (or staccatos, or accents, etc.) with a marking that has been established to signal their editorial origin.
How then, to establish the two sets of source markings, sometimes contradictory, that appear in Schubert’s viola solo part, while retaining our editorial markings (which were also necessary, for some of the reasons just described)? Here’s what we came up with for this case, as described in the appropriate section of the edition’s editorial methods, complementing our normal statements on editorial versus noneditorial markings:
Slurs added by the performer are distinguished from the original slurs by a stroke through the midpoint. The performer’s staccato markings are given as dots, while the originals are strokes. No difference in performance is intended by these markings. Note that the slurred staccato articulation also uses dots, but under a slur: this is the only case where dots do not indicate performer’s markings. Trills added by the solo viola performer are notated as a lower-case italic t. Grace notes added by the viola soloist are given diamond-shaped noteheads; unlike the scribal grace notes, the performer’s notes are generally eighth notes no matter what the context, and they are given as such in the edition. Any changes are detailed in the critical notes. Fingerings found in the original score are drawn below the staff while those added by the performer are placed above the staff.
This approach was a bit more complex than the usual treatment, but it was merited, we judged, by the specific source situation. Figures 5, 6, and 7 give passages from the solo viola part in the edition; can you see the different markings at work?
| Figure 5. Viola solo part, third movement, measures 219–23. |
 |
| Figure 6. Viola solo part, first movement, measures 275–77. |
 |
| Figure 7. Viola solo part, third movement, measures 64–68. |
 |
Hopefully, the hands of Schubert the composer, Schubert (?) the performer, Dr. Levin the editor, and A-R Editions the publisher coexist here in a score that remains legible despite its loyalty to these various progenitors. With the application of these editorial methods, born as they were from the trials of editing, we can hopefully state with Job (17:12) that “Noctem verterunt in diem, et rursum post tenebras spero lucem” (They have turned night into day, and after darkness I hope for light again).
Alexander Dean is a house editor with A-R Editions.
