Some Relationships Between the Autographs if Beethoven's Sonata in C Minor, Opus 1111 William Drabkin The complete autograph of the Piano Sonata, Opus Ill, is one of the most accessible, in reproduction, of Beethoven's manuscripts. In the last fifty years it has been published in facsimile three times: two German editions and an inexpensive American reprint.2 Unfortunately, neither German publication contains a description of the manuscript, and the introduction to the Ameri- can edition apparently ignored the existence of another autograph of the first movement of the sonata. The most significant research on the autographs of Opus III is to be found in Heinrich Schenker's pioneering edition3 and in Hubert Unverricht's critical examination of the autographs, early editions, and correspondence between Beethoven and publishers of the sonata.4 The present essay offers another examination of the autographs, particularly with respect to striking notational similarities and differences.5 I shall first enumerate the sources for the sonata, then make some observations con- cerning the general relationship between the autographs, and finally discuss some textual problems arising from a comparison of the autographs, as well as the significance of such a comparison for the understanding of Beethoven's compositional and copying processes. Sources-Autographs and Rampl's Copy The composing score, or Urschrift, of the first movement is a heavily cor- rected manuscript of seventeen pages. It is in the archives of the Beethoven- haus in Bonn, and I shall refer to it here as the Bonn autograph. The auto- graph for the complete sonata (i.e., the one published in facsimile) consists of a sixteen-page fair copy, or Reinschrift, of the first movement and a twenty- four-page Urschrift of the second movement. It is in the Deutsche Staats- bibliothek in Berlin and will be referred to as the Berlin autograph. A copy of the sonata, made for Beethoven by Wenzel Rampl, is in the Bodmer Collection of the Beethovenhaus.6 The physical properties of the manuscripts (foliation, watermarks, size of paper, etc.) are discussed in detail by Unverricht, who deduces that: (1) there is a lost Reinschrift of the second movement; (2) Rampl's copy was made from the Bonn autograph and the lost Reinschrift; and (3) an earlier copy of the Urschrift of the second movement was presumably destroyed at Beet- hoven's request.7 Sketches There are four manuscripts known to contain sketches for Opus 111:8 S. 12: an 88-page sketchbook in the Staatsbibliothek der Stiftung Preus- 38 sischer Kulturbesitz in Berlin, known as Artaria manuscript 197, or Sketch- book "D." Sketches for a fugal treatment of the main theme (mm. 21-22) of the first movement are to be found on pp. 76-77. S. 14: a l28-page sketchbook, also in the Staatsbibliothek der Stiftung Preussischer Kulturbesitz, known as Artaria manuscript 201, or Sketchbook "E." Sketches for the first movement are to be found on pp. 2-21; for the second movement, on pp. 22-62. A continuity draft of the first movement9 runs from pp. 10-21. I shall refer to this sketchbook as Artaria 201.10 S. 201: Beethoven Ms. 51 in the Bibliotheque Nationale in Paris, a col- lection of ten oblong leaves in pocket format, i.e., folded in half to form four pages of music each; these pages contain primarily sketches for Opus 110 and Opus Ill. In its present state, the manuscript is divided into nine parts, each containing one leaf (except Part 3, which has two). It is wholly unlikely that Beethoven used these ten leaves together as a continuous sketchbook, and it is quite possible that they were put together long after his death. Nevertheless, the manuscript may be reassembled into a number of "sketch gatherings," usually made up of four leaves (16 pages) which originally comprised a single sheet of paper.ll Some of the sketches for the first movement of Opus III supply missing parts of the continuity draft in Artaria 201. For example, the left-hand side of Part 8 recto has the following entry, neatly written (Ex. 1): EXAMPLE 1 Sketch for mm. 19-22. In the continuity draft the Allegro can brio begins with m. 21 of the sonata. The entry in the pocket sketchbook supplies the two previous measures. The presence of mm. 19-22 in a neat sketchbook entry helps to explain notational peculiarities which occur in this passage in the autographs. In the Bonn autograph, the printed staff lines at the bottom of p. 2 end in the middle of m. 20. Beethoven drew a continuation of the staff lines, completed m. 20, and continued to write m. 21 on the hand-written staff lines! In the Berlin autograph, he ended p. 2 with m. 19 and continued with m. 20 on p. 3, without writing the clefs and key signature at the beginning of the first system; moreover, he did not even leave enough room at the beginning of the system to write them in later.12 These notational peculiarities, which suggest that Beethoven thought ofmm. 19-22 as a unit, illustrate that even a composer's uncorrected manuscript can offer clues to his compositional process. S. 278: Ms. A48 of the Gesellschaft der Musikfreunde in Vienna, a four- page manuscript devoted primarily to a draft of the beginning of the Maestoso, in which mm. 8-10 are contracted into two measures (Ex. 2): 39 EXAMPLE 2 The Bonn autograph originally followed this version but was corrected by means of an extra measure squeezed into the above passage. Early Editions The history of the early publication of Opus III has been discussed in detail by U nverricht, Tyson, and Forbes.13 The respective dates of publica- tion of the various early editions are not known precisely,14 but the sources have been determined.15 Rampl's copy served for Schlesinger's edition, issued in Paris and Berlin. Both duplicate an error in the rhythmic notation of m. 5 of the Bonn autograph. A copy of Schlesinger's edition, together with the composer's list of errata, was used for the Vienna edition of Cappi and Diabelli.16 The complete Berlin autograph was used exclusively for Clementi's London edition.17 The General Relationship between the Autographs It is almost certain that Beethoven referred to the Bonn autograph while writing the first movement of the Berlin autograph.18 First, both manuscripts are dated in the upper left-hand corner of the first page: "am l3tenJanner (jenner) 1822." In all probability the Bonn autograph was begun on this date, since the last page of the continuity draft of the first movement (Artaria 201, p. 21) bears the inscription: "am 13ten die neue Sonate." Second, the number of pages in both manuscripts is the same, and a majority of cor- responding pages begin with the same measure.19 Third, p. 8 of both auto- graphs lacks clefs and a key signature at the beginning of the first brace; in both manuscripts, however, there is ample space for them. Finally, at the top ofp. 12 of the Berlin autograph, the left-hand part is corrected to read an octave higher. This change is the result of Beethoven's copying the beginning of p. 12 of the Bonn autograph without realizing that the "8va" on p. 11 thereof is still operative after the page turn. When Beethoven wrote the Berlin autograph, he avoided using the "8va" on p. 11. He was thus obliged to correct the notation of the left-hand part on p. 12 for the sake of uniformity. The first movement of the Berlin autograph is a Reinschrift insofar as it contains no compositional corrections. Page 3 has a "Vi-de" cancellation of a measure which was written with poor alignment between the right- and left-hand parts; similarly, a few notes on p. 13 are crossed out in order to 40 improve the alignment of parts. I have already mentioned the corrected copying error at the beginning of p. 12. Although there are differences in content between the autographs of the first movement, none is the result of a correction in the Berlin autograph. Variants in the Berlin manuscript consist of different note readings in two measures and additions and omissions of dynamic markings. The variants in dynamic markings are discussed in detail in Schenker's Erlauterungs-Ausgabe, passim. The significance of the different note readings will be discussed later in the present essay. Some Specific Relationships between the Autographs-Key Signatures The absence of clefs and a key signature on p. 8 of both autographs raises the following textual question: Does the key signature at the top of p. 9 (m. 87 in both manuscripts) indicate a change of signature, or is there a missing signature change before m. 87? Although Beethoven customarily wrote out the clefs and key signature for the first system of each page of his piano sonata autographs, Rampl con- strued the signature on p. 9 of the Bonn autograph as a change of signature, and Beethoven did not correct Rampl's copy in this detail. All the editions of the sonata except Clementi's indicate a signature change at m. 87; in Cle- menti's edition, a three-flat signature replaces one of two flats at the begin- ning of the fourth system on p. 6 (m. 86). I should like to reconsider the validity of this signature change: an ex- amination of the notation of p. 8 of both autographs reveals that the change of signature ought to have occurred much sooner-perhaps as early as the beginning of p. 8, precisely where Beethoven left room for a key signature. Let us assume, however, a two-flat signature on p. 8 of the Bonn autograph. Then mm. 77, 85, and 86 have A's with superfluous natural signs, and mm. 81-84 (and the right-hand part of 80) are textually incorrect, since there are no flats in front of the A's in these measures. Of the remaining measures on p. 8, 76 contains no A's, and 78 and 80 (left-hand part) contain A's with flats that cancel previous natural signs. Thus, from a textual point of view, only m. 79 can be more appropriately considered in a signature of two flats-the flat in front of the right-hand A does not cancel a previous natural sign. But this flat was almost certainly added to the autograph after Rampl copied the sonata, since it is missing from his copy. One may therefore conclude that Beethoven had in mind a signature of three flats when he wrote p. 8 of the Bonn autograph.20 The accidentals on p. 8 of the Berlin autograph are essentially the same as those in the Bonn autograph; this situation suggests that Beethoven copied it mechanically and hastily. There are flats in front of the A's in the right-hand part ofmm. 79 and 81, but none in m. 80.21 If Beethoven intended a signature of three flats on p. 8, why did he bother to make a previous change to two flats for a phrase only four measures long (72-75)? The purpose of such a signature change is difficult to explain, but a 41 precedent for it is found in the first movement of the Sonata in E major, Opus 109: the signature change at m. 61 is cancelled at m. 63. Beethoven was also uncertain of the change of key signature at m. 72. In the Bonn autograph, this change originally occurred at the beginning of m. 69a (the second ending)-presumably the result of the composer's absent- mindedness. (The octave are intended here, as in the first ending!) Rampl's copy and Schlesinger's edition preserve the error. When Beethoven discovered his mistake, he crossed out the signature change in 69a and placed it at the beginning of 72. In the Berlin autograph Beethoven copied the rests at the beginning of m. 72 before writing in the signature change. Moreover, in both autographs the previous measure (71) is textually incor- rect in the old signature of four flats, since there are no natural signs in front of the right-hand D'S.22 In sum, Beethoven's intended key-signature changes, as evidenced by the accidentals in the measures preceding 72 and 87, are represented accurately neither by the signatures found in the autographs nor by those which appear in early and modern editions. Variants in Note Readings There are only two differences between the autographs with respect to note content (excluding differences which were corrected in the early editions): (1) the last left-hand chord of m. 116, and (2) the penultimate sixteenth-note in m. 133 (Ex. 3): EXAMPLE 3 a. M. 116 Bonn Berlin b. M. 133 (last quarter). Bonn Berlin The Bonn versions appear in every edition which I have consulted (except, of course, Clementi's early edition). The only account of the variant readings is to be found in the Erlauterungs-Ausgabe, but Schenker does not even men- tion the discrepancy between the versions of m. 116.23 It is possible to demonstrate that the readings in the Berlin autograph represent stages in the composition of the work that precede the corrected version of the Bonn autograph. In other words, we have a paradoxical example of a manuscript which was written later representing the earlier stage of the composition, at least in the two places in question. 42 Measure 116: There are several corrections of notes and tempo markings in Rampl's copy which correspond to identical corrections made by Beet- hoven in the Bonn autograph; these were the last compositional changes made in the first movement.24 Among them are: (1) the addition of a quarter-note d in m. 77 (and the consequent restemming of the left-hand part); (2) the changing of the left-hand part ofm. 91; and (3) the changing of the last left-hand chord in m. 116, which originally read: g-c1-el, the same as the Berlin version. The new readings of (1) and (2) incorporated in the Berlin autograph support Unverricht's hypothesis that Beethoven wrote the first movement of the Berlin autograph some time after Rampl's copy had been corrected and sent to the publisher in April 1822; had he written it earlier, the Berlin autograph would also have had corrections in these places.25 If the three corrections were made at about the same time, then the Berlin version of m. 116 can be ascribed to a mental lapse on Beethoven's part, during which he wrote down the earlier version.26 The Bonn version ofm. 116 agrees with the parallel place in the exposition of the movement, m. 50. The chord here is eJ,-cLeJ,l, and there are no cor- rections in this measure in either manuscript. One may be inclined to ask: Why does the chord become problematic in the recapitulation, and not in the exposition; is the Berlin version actually an error, or was the composer perhaps attempting to vary the recapitulation? I do not believe that Beet- hoven, having discovered an ingenious disposition of the I2-chord (in which the right-hand melody supplies the root of the chord two octaves higher, and in which the right- and left-hand parts are enabled to outline the same EJ,-DJ,-C motive-see Ex. 4), would have abandoned it in the recapitula- tion for the sake of variety. EXAMPLE 4 X X X t. ......... t: .e #- 7 sf - f X X X (51) :t- -&-: ,::: * Mm. 50-52. The X's delineate the motive in the right- and left-hand parts. In the continuity draft in Artaria 201, m. 116 is better defined and closer to its final version than m. 50 (as may be observed in Ex. 5 on following page). Noteworthy features of the sketch of 116 are: (1) the right-hand melody begins on the second beat, instead of the first; (2) the large numbers under the melody explicitly indicate two left-hand chords, instead of an arpeggio; and (3) one of these chords, a 11, is shifted to the fourth beat of the measure. It appears that Beethoven had not yet decided to delay the I1-chord until 43 EXAMPLE 5 M.50 8 M.1l6 t the final sixteenth of the measure. Since the right-hand c3 does not occur until the final sixteenth, Beethoven needed a C in the left hand in order to have a complete chord; thus he probably had in mind an ordinary left-hand chord for m. 116 in the continuity draft. For the evolution of m. 116 I suggest the following scheme, based on the corrections in Artaria 201 and the Bonn autograph (Ex. 6) : EXAMPLE 6 Stage 1 Stage 2 Stage 3 Stage 4 Note that the last chord is always a complete but is achieved in the final version alone by the right hand complementing the left. It must also be made clear that Beethoven reached Stage 4 in manuscript bifore Stage 3, inasmuch as he wrote m. 50 of the Bonn autograph before he wrote m. 116; thus Stage 3, which combines elements of Stages 2 and 4, represents a backwards develop- ment in the composition ofm. 116. In conclusion, I offer the following explanation for the existence of the error in m. 116 of the Berlin autograph and the absence of such an error or cor- rection at m. 50: There is a continuous written evolution of m. 116 (illus- trated in Ex. 5), in which the error is an intermediate step. Since no such evolution exists for m. 50, the error never shows up in the autograph material. Measure 133: The discrepancy between the versions of m. 133 may also be traced to an earlier stage in the Bonn autograph. After trying a number of unsatisfactory solutions for mm. 132-34, Beethoven crossed out all the un- usable material for these measures and worked out the final version on a sep- arate page.27 This page contains more rejected solutions, but one of them was 44 close enough to the final version to permit Beethoven to indicate the final version merely by changing a few notes in the rejected one (Ex. 7): EXAMPLE 7 a. First version h. Final (corrected) version M. 133 in the Bonn autograph. The penultimate sixteenth-note of m. 133 is a G in Example 7a, illustrating that the reading in the Berlin autograph is part of a rejected version in the Bonn autograph. Hence, the errors in the Berlin autograph at mm. 116 and 133 are attributable to the same cause: the mental copying of a stage pre- ceding the final version of the composition. In view of the accessibility of the Berlin autograph, it is regrettable that the Bonn autograph has not also been published in facsimile. The presence of wrong notes in the Berlin autograph illustrates once more the danger in- volved in working with only part of the existing source material for a com- position. In accounting for variants, I have illustrated only a few of the corrections in the Bonn autograph. But an exhaustive study of all the correc- tions in this manuscript-and of the related sketches-would help us better understand Beethoven's procedures in composing the first movement of Opus 111.28 NOTES 1 This paper was based on a project for a seminar on Beethoven conducted by Professor Lewis Lockwood at Princeton University. 2 Drei Masken-Verlag (Munich, 1922); C. F. Peters (Leipzig, 1952); Dover Books (New York, 1968). The Dover edition, in paperback, is a reprint (and reduction in size) of the Peters edition, which, in turn, reprints (and cleans the page borders of) the edition of Drei Masken-Verlag, made on the IOOth anniversary of the composition of the sonata. Although the Dover edition is reduced in size and printed on cheaper paper, pages which have few or no corrections can be read with little more difficulty than either German publication, except 45 for editorial accidentals in pencil and not in Beethoven's handwriting. Heavily corrected pages (in the second movement) are easier to read in the larger German publications; how- ever, it is difficult to determine the original readings of these passages in any of the facsimile reproductions. 3 Published in 1915 as part of Die letzten funf Sonaten von Beethoven, mit Einfuhrung und Erlauterung von Heinrich Schenker (Vienna, 1913-21), commonly referred to as the Erliiuterungs- Ausgabe. 4 Die Eigenschriften ulld die Originalausgaben von Werken Beethovens in ihrer Bedeutung fur die moderne Textkritik (Kassel, 1960). 6 I shall also discuss some relationships between sketch and autograph and, as they concern peculiarities in the autographs, between sketch and sketch. 6 Described in Max Unger, Eine schweizer Beethovensammlung (Zurich, 1939), pp. 158-59. In his article "Notes on Five of Beethoven's Copyists," JAMS (1970) 23: 438-71, Alan Tyson challenges the attribution to Wenzel Schlemmer of this and other copies of Beethoven's late works. Unverricht follows Unger in attributing this copy to Schlemmer, who was thought to have been Beethoven's chief copyist from about 1807 until his death in 1823. But Tyson demonstrates convincingly that it was made by Beethoven's "Copyist B," who he believes was Wenzel Rampl (pp.450-52). I am indebted to Dr. Hans Schmidt of the Beethoven-Archiv for providing me with a microfilm of this copy. 7 Unverricht, op cit., pp. 34-38. See also Emily Anderson, The Letters of Beethoven (New York, 1961), Volume II, p. 942 (letter 1074). 8 The most extensive published catalogue of Beethoven's sketches is Hans Schmidt's "Verzeichnis der Skizzen Beethovens," Beethoven-Jahrbuch 1965/68 (Bonn, 1969), pp. 7-128 (entry numbers prefixed with "S." in the present essay). 9 On the use of the term "continuity draft" see Lewis Lockwood, "On Beethoven's Sketches and Autographs: Some Problems of Definition and Interpretation," Acta Musicologica (1970) 42:42. 10 Both Artaria 197 and 201 are described in Gustav Nottebohm, Zweite Beethoveniana (Leipzig, 1887), pp. 468-75. 11 Parts 4, 6, 8, and 9 once comprised a single sheet of paper, and there are a number of musical connections between pairs of pages, which suggest that they were used as a 16-page pocket sketch gathering for the last movement of Opus 110 and for both movements of Opus III. Similarly, Parts 1, 2 and 5 were once part of a single sheet, the fourth corner of which is not to be found in this manuscript. These leaves, with or without the missing leaf, probably also comprised a sketch gathering for exactly the same music. In both of the above recon- structions, all the sketches for Opus 110 precede those for Opus III, and there are no musical connections between the gatherings (although they may well have been used at the same time). There are no connections between these gatherings and the remaining parts of the manu- script: Part 7 contains sketches for the Agnus Dei of the Missa Solemnis, as well as for both movements of Opus Ill; Part 3 consists of an entirely different kind of paper, which, together with Ms. 80 and Ms. 99 of the Bibliotheque Nationale, comprises a 24-page pocket gathering containing sketches for the Credo of the Missa Solemnis and all three movements of Opus 110, as well as a number of drafts of the canon for Tobias Haslinger, WoO 182. 12 Beethoven supplied the clefs and key signature for the first brace of every page of the first movement of the Berlin autograph except p. 8, but there he left sufficient room to write them in later. 13 Unverricht, op. cit., pp. 34-47; Alan Tyson, The Authentic English Editions of Beethoven (London, 1963), pp. 110-13; Thayer's Life of Beethoven, revised and edited by Elliot Forbes (Princeton, 1964), Volume II, pp. 859-61. 14 Recently Dr. Tyson informed me that in his opinion all of the early editions of Opus III date from the spring (or early summer) of 1823. 16 Unverricht (op. cit., p. 34) summarizes the principles used in determining the relation- ship among the autographs and early editions: "Die Abhangigkeit von Quellen lasst sich an 46 Hand von gleichbleibenden Behalsungen und Balkungen, durch ilbernommene Fehler ... und Varianten, an gemeinsamen weggelassenen oder erganzten Akzidenzien und Vortrags- zeichen, gelegentlich durch genaue Ubereinstimmung der Zeileneinteilungen und anderem nachweisen." 16 The list of errata is included in Anderson, op. cit., Volume III, p. 1047 (letter 1I90a). 17 See Tyson, Authentic English Editions, pp. 110-13, for a list of variants among the early editions. 18 Unverricht (op. cit., p. 38) asserts: "Beethoven hat die Reinschrift nicht wieder von neuem aus dem Kopf geschrieben, sondern die Urschrift kopiert, wie auch aus den vielen gleichzeitigen Bemerkungen in der Erstschrift und aus der oftmals getreuen Einhaltung der gleichen Zeileneinteilung hervorgeht." The material I have gathered here documents Unverricht's claim. 19 The Bonn autograph is actually a page longer, but the last page consists of an insertion for a heavily corrected passage p. 14. (See the discussion ofm. 133, towards the end of this essay.) 20 Most of the textual errors in mm. 79-84 were corrected in Schlesinger's edition by flatting some of the A's in these measures. It would probably have been impossible to change the position of the key signature without reengraving the entire first line ofp. 7. In my opinion there is no musical rationale for a key-signature change at m. 87, the second measure of the phrase which leads to the recapitulation of the movement. Thus Clementi's solution, which seems to be quite unintentional, is more reasonable from a musical point of view. 21 Curiously, mm. 82-83 in the Berlin autograph must be read textually in a signature of four flats (F minor, the local key center of these measures), since there are no flats in front of the right-hand D's. Clementi's edition preserves all of the textual errors in mm. 80-84 of the Berlin autograph. 22 Apparently Beethoven failed to take into account his having shifted the signature change from m. 69a to m. 72. The edition of Cappi and Diabelli indicates the change at m. 71, despite Beethoven's admonition in the Bonn autograph. In failing to heed Beethoven's instructions, it thus corrected the textual error in m. 71. 23 Furthermore, he treats the Berlin version ofm. 133 as an oversight on Beethoven's part (p. 48): "Sonderbarweise zeigt Aut. b) [Berlin] als vorletztes 16tel des T. 133 den Ton g statt as; kein Zweifel, aber, dass wir hier nur mit einem Versehen zu tun haben." 24 Later editorial changes, such as the insertion of accidentals and the shifting of the key signature from m. 69a to m. 72, are not to be considered compositional; neither are the dif- ferent readings of m. 116 and m. 133, as I demonstrate below. 25 See Unverricht, op. cit., pp. 36-37. Beethoven probably wrote the first movement of the Berlin autograph while engaged in negotiations regarding an English edition of the sonata. These negotiations began inJuly 1822, and Beethoven sent the Berlin autograph to Ferdinand Ries in London in February 1823. See Anderson, op. cit., Volume II, p. 954 (letter 1084), and Volume III, p. 1006 (letter 1143). 26 It is also possible that Beethoven changed m. 116 much later, after he had written the Berlin autograph, in which case the Berlin version would have been the result of his copying the Bonn autograph directly. But this possibility seems unlikely, since the correction in m. 116 would then be the only compositional change in the first movement made after its Reinschrift was written. 27 See fn. 17. Beethoven also had difficulty composing a similar passage, mm. 56-57, on p. 6 of the autograph. The sketches on pp. 12 and 19 of Artaria 201 show additional unsuccessful attempts to find a satisfactory register and descending motivic pattern for mm. 56 and 132-33. 28 A microfilm of the Bonn autograph is in the Toscanini Memorial Archives of the Music Division of the New York Public Library at Lincoln Center. The sketches are on film in the Princeton University Library. 47