Manuscript Sources


   Comparison of Durand edition with Biblioteque National manuscript, Winterthur manuscript; parallels with violin sonata and proof. Deductions as to the differences between manuscripts and published edition.






Two  manuscripts of the Sonata exist.  Comparison of the two  - one in the Bibliothèque Nationale, Paris, and one in the Winterthur Stadtbibliothek  ( from now on referred to as the BN Ms and WSB Ms) - reveals some intriguing differences between those and the Durand edition, which was based on the BN manuscript. The study of Debussy's manuscripts reveals great precision and clarity in the draughtsmanship. Everything is laid out with meticulous care in such a way that it has very much the same overall appearance as the printed edition, though the note sizes are much smaller and pointillistic, executed with a fountain pen in blue ink.


The WSB manuscript is a fascinating insight into the process of composition, as it consists of Debussy's “ working ” copy of the sonata, with other sketches, which include three versions of the opening of the first movement.  At this stage there are no markings for dynamics, expression or articulation, just the notes, and we can see the evolution of the piece taking place as various ideas are coaxed into shape.


Debussy's conception for the opening piano bars remained basically the same in all the versions, with the exception of the rhythm immediately after the first chord, which was a much less concise and characterful version of the ( final ) demi-semiquaver triplet motive, but the cello entry underwent various transformations before the final shape emerged. ( Example 1 )



Ex 1  First Movement Sketches from the Winterthur Manuscript




The first versions play around with the same melismatic idea, but by comparison with the final version, also seem much less focussed and concise. The melodic material has an organic growth from one idea to another, with considerable play around certain melodic cells, and groupings of notes. One can really feel the way in which Debussy was exploring this material as he went along, so that the structure evolved as a result of ever-renewing variation, with one section added to another to create the whole.


At the top of  the first page of the second movement, there is an isolated phrase  jotted down next to the cello pizzicato theme from the second movement ( Example 2 ) which is one of those cells, in a rough hewn version, written out in such a way that it looks rather like a plainsong melody.  This whole tone sequence of the descending fourth from E to Bb recurs in various places in the sonata, as does its outline interval of a tritone.





Ex 2   "Tone Row " from second Movement - Winterthur Manuscript - in Debussy's hand


( See Footnote No 2 )


It is interesting to compare bars 8 and 9 of the WBS manuscript Sérénade with the BN manuscript and the Durand edition, ( Example 3 ) as all three are different. In both manuscripts Debussy wrote a glissando - the second  (where the words glissando chromatique  are written above the music ) even more explicit than the first - but in the Durand edition this  powerful and emotive gesture has gone. Was Debussy responsible for  its removal at proof stage?  On reflection, it seems  that that the published version is the safest, but least interesting version of the three, and it seems a pity that after it was written in both manuscripts, it did not appear in the Durand edition. 








Example 3


The second and third movements were a little less straightforward in the composition, though basically the conception was clear and the flow of ideas, apart from a few crossed out  and slightly reordered sections, was unbroken.  It was more a case of reworking  and refining, rather than rejecting, ideas, as can be seen from the WSB  manuscript, and the general impression one gets from this first manuscript is of a great rush of ideas, spilling onto the paper with incredible fluency, uninterrupted from beginning to end.


Two small points of interest: six bars before 8 ( in Durand ), the WSB manuscript has the description “ sons harmoniques ” and  this became        “ sur le chevalet ”  in the BN manuscript ( which sheds some light on the Flautendo marking in the Sérénade - see Part Three ). Then later on, four bars just before the Lento at Fig 9, Debussy has written below the piano part


            “ Que le pianist n'oublie jamais qu'il ne faut pas entrer contre le  violoncelle mais l'accompagner  ” [1]


 - a useful homily to pianists who often  get carried away with enthusiasm at that point, and an unusually far-sighted comment, considering the fact that the piece was still very much in the evolutionary stage, and that this was in every other respect a sketch.


Looking next at the BN manuscript, we can see the composer's immaculately detailed fair copy ready for the engraver  Charles Douin, with some pencilled instructions ( in a hand other than Debussy's, probably that of Jacques Durand, ) indicating the format for the engraving of the title page, and  adding  “ La partie du violoncelle sera extraite ulterierement ”.


It is extremely revealing to compare the BN  Violin Sonata manuscript and first proof, with the BN manuscript of the Cello Sonata, since, in trying to establish whether the differences in the BN Cello Sonata manuscript and the Durand edition occurred at proof stage  ( since no corrected proof of the Cello Sonata exists ), Debussy's practice in the case of the Violin Sonata may give us some clues. What is very noticeable is that a great number of changes were made in the Violin Sonata between composition and publication, and that Debussy himself, when correcting the proofs, tended to add  ( rather than delete ) many more markings in dynamics, tempi and articulation than had even been in his final fair copy.


But it is also significant is that whereas the fair copy BN manuscript of the Cello Sonata is absolutely clean, without a single note changed in the Durand Edition, the Violin Sonata fair copy manuscript has many crossings-out and changes. This would corroborate the accounts of the difficulties that Debussy encountered in the composition of the Violin Sonata, and particularly the last movement, which in the fair copy has  another beginning consisting of several bars of quite different music crossed out before the final version appears. ( By all accounts, the writing of the Cello Sonata seems to have caused Debussy far fewer  problems than that of the Violin Sonata. )


These last minute changes indicate quite considerabe uncertainty on Debussy's part, and necessitated the large number of corrections in the proofs. This is not the case with the Cello Sonata, where many of the differences seem much more likely to have been  omissions by the engraver, differences of “ interpretation ” of the manuscript, such as inconsistencies occuring  in the original, or simply mistakes. The question remains, did Debussy make the changes himself ? Unfortunately in many cases  the matter cannot be conclusively resolved.


After publication, Debussy's remark to André Caplet “ As for printing mistakes, contact Jacques Durand ”   suggests that either he was unaware of any himself - assuming that he had seen the published edition  - or that he was not particularly concerned with the matter anyway, and left it to Durand to sort out. Effectively he gave over the responsibility for further corrections to Durand, which might also explain the  existance of differences between fair copy and printed edition.


Looking at the  manuscript, one has a very strong impression of the clarity of the sound world conceived in the Sonata, as though every note and articulation occupies an exact  place in the composition as whole. And there are also some aspects of the Durand edition, which, though not mistakes, nevertheless affect the look of the music in a way which is not completely sympathetic to Debussy's manuscript, which in so many ways is exemplary.


For example, Durand had a habit of inverting the accent signs ^ , ( changed to v ), and often puting them underneath the notes to which they refer, which Debussy never did. The sign when inverted also looks confusingly like an up-bow sign. The way in which Debussy wrote Tempi instructions gives some indication of those which really were Tempi, as opposed to modifications of those Tempi, such as the frequent Cédez markings, often written in a less strong script.


At the beginning of the Sérénade, in the piano score of the Durand edition the instruction Modérément animé appears above the piano part, with          “ fantasque et léger ”  below the cello part. In the Ms however, the two things are written on the same line in the same strength above the stave, which suggests rather a different thing. ( the cello part is correctly marked ).


There are no metronome marks on the Ms, just empty brackets. These must have been a very late addition.


There are so many small details in which the Durand edition differs from the BN manuscript that they are too numerous to point out here, the majority being articulations, such as dots, slurs and dynamics. A Peters Edition of 1969, though it does list the major changes, is also not a facsimile  of the original.


The following is a list of some of those differences between the BN Ms and the Durand edition, omitting those minor slips where Debussy himself omitted to repeat articulation. It is not intended as an errata list, but simply rather as an observation of the changes.







Bar       1          Pno - no stress marks on last four notes.

            2          Pno - no cresc

            3          Pno - slur over last beat in RH upper part

            5          Pno - cresc over first beat; sfz on second quaver chord

            15         Pno - hairpin down as cello

            22         Cello - pp mark at beginning of  bar

            31         Cello - no slur over last beat

            32         Pno - hairpin down instead of cresc

            37         Cello - no dynamics except first hairpin

            38         Cello - no molto dim: a hairpin up to the middle of the bar

                        and a hairpin down to end of bar




Bar       6          Pno - slur over last beat on to first chord in next bar

            7          Pno - last chord Lh spread as RH

            8          Cello - chromatic gliss from harmonic to last beat

            9          Cello -                                    "

            22         Cello - gliss between the G and Ab

            39         Cello - slur above the whole bar

            41         Cello - dots above first two semiquavers and no ƒ

            41         Pno - no ƒ

            49         Cello - slur over pizz's ( also in 51 with no cresc )

            51         Cello - hairpin down over first two quavers ( as Pno )

            52-3      Cello -  slur over both bars and over 53 into 54, indicating

                        the resonance.

            61         Pno - P as Cello





Bar       18         Cello - no Dim

            18         Pno - dots over LH quavers and no slurs

            29         Cello - slur over bar line from last E

            31         Pno - Dim over first group on demi-semi's

            49         Cello - no cresc markings

            51         Cello and Pno - no cresc markings

            55         Cello - last semi's separate

            57         Pno - slurs under LH part go from bar line to bar line - 6 bars





[1] “ ...... the pianist should never forget that he must not enter against the cellist, but must accompany him. ”




Following publication of these articles in the Strad Magazine of April - June 1992, two letters were received  ( the first in Part One ) which throw more light on points discussed. "Debussy's Lutheran Side ? " should be read in conjunction with the "Tone Row" discussion above.