Gaspard Le RouxGaspard Le RouxGaspard Le RouxGaspard Le Roux

The Sources

The Sources

Only three complete sources of Pieces de Clavessin remain: the aforementioned copies found in the Bibliothèque nationale de France, Library of Congress, Washington and the Fitzwilliam Museum, Cambridge. Of these, the Fitzwilliam Museum houses the only known Estienne Roger edition.[1] A copy of his second harpsichord part is also found in the Berlin Staatsbibliothek, as is J G Walther’s manuscript copy of about half the contents of the Paris imprint. The only previous modern edition was prepared by Albert Fuller and published in 1959. This is intelligently handled, with interesting prefatory information and a thoughtful commentary. Fuller has correctly relied on the Paris imprint, using the Washington copy as his primary source, and only occasionally does he refer to the Cambridge Amsterdam copy, which he calls a ‘slavish’ version of the original.[2] More recently, Colin Tilney reproduced the preludes as part of a three-volume edition that contains insightful corrections, facsimiles of the 1705 preludes, modern transcriptions and a brief commentary.[3] In addition, individual movements from the pièces have found their way into a number of anthologies, but these are too numerous to list here and are of no consequence.

The 1705 Paris Edition

This is a clear engraving and although it contains considerable inconsistencies, its text is largely free from glaring errors. Printed in quarto landscape format, it presents the harpsichord pieces in combination with their trio arrangements, which occupy the lower portion of each page. The music is rather cramped and it does not have the look of the usual work that came from Baussen’s atelier, which may be seen at its best in Courante luthée (I) and Double de la Courante (VI), where the trios are omitted. Why these dances have no arrangements warrants some discussion, since they do not correspond with the overall plan adopted by Le Roux. It could be that the brisure of Courante luthée made it unsuitable material for other instruments, or that he envisaged its performance only when approaching the suites as solo works. This is improbable, however, since not only does it counter Le Roux’s philosophy, but it also disregards the fact that similar textures were arranged elsewhere.[4] In addition, since Le Roux saw the trios’ contreparties as essential when performing the dances as duets, their inclusion would have been desirable, if not expected. Also, it does not explain why the simpler texture of the double was left unarranged: space was not an issue, as the inclusion of the extra trio of either dance would not have exceeded the space allotted its solo version and engraving costs would not have been prohibitively more expensive. It could be, though, that they were last minute substitutions for movements that Le Roux, for some reason, decided to remove. What they might have been is a matter of conjecture: the first suite follows convention and ends with a quicker dance, so it is possible that Courante luthée replaced a prelude to Suite II; the double might have been used in place of another sarabande, a different courante or, as will be discussed, the loure La favoritte. To speculate further, the existence of the solo engravings could indicate that Le Roux’s first plan was different to the one we know, his intention being to publish the harpsichord works alone. If this were the case, then it is probable that work had begun on engraving before a choice was taken to enlarge the pièces to the format that has come down to us. A further decision to change some of the content might then account for the unarranged movements, since by that point it could have been more efficient to use the earlier plates.

Although first impressions suggest something similar happened with Sarabande (VII), a more thorough examination of the music shows a clever and radical solution to what must have been a problematic piece to engrave and print. Since Le Roux draws particular attention to the dance in his preface, it is fair to assume that it was something of which he was proud. Its trio arrangement, however, seems unfinished: after the first couplet both treble instruments are tacet, leaving the second, third and fourth to be accompanied by basso continuo alone. The remaining eight are scored only for the harpsichord. Yet it seems that this is by design, with the sarabande’s organisation allowing several approaches in one. The trio might be played as it appears, a single movement followed by the suite’s final minuet. The second, third and fourth couplets could indicate that the right hand of the harpsichord part was to be taken by a solo instrument. This would account for Le Roux switching to the otherwise unnecessary G1 clef in the second couplet, making it easier for an instrumental soloist to read. From the fifth couplet onwards the harpsichord itself becomes the solo instrument. It could be, therefore, that Le Roux was maximising the potential of what would have otherwise required more space than its eleven pages employ, which would have increased production costs considerably.

One other problematic piece is found at the end of Suite VI. La favoritte has the wrong page numbers of 55 and 56, which are already taken up by Double de la Courante and its following Sarabande Grave. Not only is the numbering incorrect, but it also appears twice on each page in opposing corners of the top margin. Although this was evidently a correction on the part of the engraver, who began the movement on the recto rather than the preferred verso side, it is necessary to question how the numbering became so confused. It may be agreed that the dance was intrinsic to the book, since it comes with a trio arrangement. Yet there is something intriguing about its content that sets it aside from the other pièces as it requires a subtler interpretation, especially in some of its difficult compound ornaments. It could be that, on realising this was the case, Le Roux removed it altogether, perhaps even replacing it with the double and the sarabande, only to return it at a later point, which would then explain the non sequitur in the page numbering. Why that happened might be to do with the length of the suite, which would otherwise have been quite meagre, but it has been reasonably argued by Tuffery that the title of the work is a reference to Madame de Maintenon, the erstwhile mistress and later secret wife of Louis XIV. As such, Le Roux might have thought it an effective means of courting favour for the publication.[5]

Both the Paris and Washington copies have considerable emendations in the form of additions, corrections and deletions, most of which have been added by the same hand: the hologram is the same and, in turn, it matches the appearance of the engravings. The revisions have been made with the same precision and, generally, no trace of the original content is apparent. It is only when comparing the solo parts with their trio versions that the extent of the alterations may be understood. Most cover minor engraving issues such as wrong accidentals, slurs and ornaments, but there are instances where musical material has been added, or where titles have been provided for pieces that previously lacked them. Most of the same corrections occur in Estienne Roger’s 1708 edition, as well in Johann Gottfried Walther’s manuscript copy (hereafter P 801), with the exception of the title given to the third movement of Suite I.[6] Presumably, it was Walther who thought up the appellation Air since none was provided in the engraving: both printed versions have the handwritten title Sarabande Grave and this was clearly an emendation that had not made it to Walther’s source. Yet another handwritten addition that appears in both the Paris and Washington copies is a title that was appended to the allemande of Suite III: l’Incomparable. These emendations also appear in Walther’s manuscript and Roger’s edition, providing a link between the corrections and suggesting they have the same provenance. Similarly, the uniform approach found in the two surviving 1705 copies suggest they were batch corrections that could have been administered by Le Roux himself.

Thus, we have a somewhat problematic source that raises a number of questions concerning its genesis. While little of this is relevant to the music itself, the lack of concrete information concerning Le Roux’s life requires a more thorough examination of the philosophies that influenced Pieces de Clavessin. The combination of solo pieces with trio arrangements as well as the realised contreparties suggests Le Roux’s vision was not purely one of aesthetics. Its commercial appeal is one of the publication’s most notable features and while it might be expected that today’s entrepreneurs capitalise on their respective markets, this is something that is rarely considered from the viewpoint of an 18th-century Parisian musician. From this perspective alone, Le Roux’s pièces document an aspect of music history that has yet to be researched fully.

The Amsterdam Edition and D-B Mus. ms. Bach P 801

Flourishing between 1696 and the year of his death 25 years later, Estienne Roger came to be one of the most important influences in European music publishing, taking the dissemination of printed scores from a parochial level to one of cosmopolitan proportions. His distribution network was supported by agents in many European cities, with representatives in both London and Cologne as early as 1706. By 1716, his publishing empire had expanded to include Rotterdam, Berlin, Leipzig, Hamburg, Halle, Liège, Brussels and Paris. At the time of his death, Roger’s catalogue numbered some 600 works, and although reprints of Italian instrumental music and French airs made up the bulk of his early inventory, by 1710 it included such leading composers as Corelli, Albinoni and Vivaldi.[7] Among the harpsichord composers represented were Lebègue (Premier and Deuxième Livres, 1697 and 1698), Marchand (Premier and Deuxième Livres, 1701 and 1703), Dieupart (1702) and D’Anglebert (1704). A further addition was made to the 1708 catalogue: ‘Pieces à 1 & 2 Clavessins, composées par Mr. le Roux. avec la maniere de les joüer´. Roger prided himself on both the quality and inexpensiveness of his editions, and its six-guilder price was approximately half that of Le Roux’s imprint.[8] Subsequent catalogues indicate that the suites were available until 1743, when the last owner of the Roger publishing house, Michel-Charles le Cène, died.[9]

Published in upright folio format, Roger’s imprint was provided in two separate parts, with the realised contrepartie examples forming its five-page supplementary volume. Since the title page emphatically points out ‘Propres à joüer Sur Un & Deux clavessins’, it must be assumed that the volume’s versatility was something Roger was keen to exploit. Yet the somewhat paradoxical omission of the trio arrangements, of which the contreparties were to facilitate performance by those not wanting to improvise a secondary part, could indicate that Roger was unaware of the true nature and scope of the original.

Fuller and McKean are quick to dismiss Roger’s edition as little other than a ‘slavish’ reproduction of the original, yet this is somewhat misleading since it implies a level of inferiority that makes the volume unworthy of consideration as a source. To the contrary, its production was undertaken by an experienced engraver, is largely true to the original and although harmonic errors were to remain uncorrected, it nevertheless underwent a process of revision during its engraving.[10] For the most part, this was the application of Roger’s house style, which provided uniformity in its notational conventions: the preferred F3 clef is altered to F4 and the seemingly haphazard approach to beaming standardised. An effort to address some of the ambiguities found in the original edition is also evident, and while not having the aesthetic appeal of Baussen’s work, it is a clean and accurate version that is representative of the reputation for which Roger was famous.[11] Aspects of the 1705 imprint also remain intact, including the retention of both original stem directions to illustrate note leading and liaisons to show which notes should be held longer than their allotted values.

However small Roger’s alterations might appear, they remain significant as they reflect notational innovations that were to become the European standard and for which his press might have been a contributing factor. It also exemplifies a growing shift towards the dissemination of musical tastes across national boundaries, which might not have been as effectively achieved without a printed media that employed a standard notational orthography. From this perspective, the debate concerning his somewhat cynical business practices, which saw the wholesale copying of other publishers’ works, is of little relevance. Roger offered an international repertory to a wide enough audience for Le Roux’s music to be mentioned in Walther’s Musicalisches Lexicon, where both entries concerning the composer bear the Amsterdam publisher’s name.

It is somewhat incongruous, therefore, that Walther’s c.1714-1717 copies of three of the suites were based on the Paris imprint.[12] As mentioned, it contains the same emendations, the dances from Suites I and III are prefaced with a reproduction of the edition’s title page, and the predominant clef used for the left hand is F3. While largely accurate, Walther left out a considerable number of ties, ornaments and notes, and seems not to have grasped the function of either the liaisons, which he often mistakes for slurs, or the written out coulés, where he fails to understand their significance by omitting their slurs altogether.[13] Despite such problems, his copy still contains a couple of revisions that come in the guise of corrected parallel harmonies and the rescoring of awkward passagework.[14]

The remaining works, which comprise most of Suite VII, are not grouped with Suites I and III but found in another part of the manuscript. Again, the F3 clef is mainly used for the left hand, which informs us that it was based on the 1705 edition. However, a degree of confusion concerning the composer’s name indicates that his source was not the printed version, but a manuscript copy (hereafter, Source Y). Its title page (423) provides what appears to be a portmanteau of Clérambault and Le Roux – Clerambouloux – which McKean plausibly suggests was a misreading of the composer’s full name.[15] While he is most likely correct, a second attribution that reads ‘Clerambaut’, which is found in the top margin of the prelude, indicates that Le Roux’s music was unknown to Walther at that time, and the suite must, therefore, have been copied at an earlier point than the others. In addition, Walther added a short inscription at the end of the final minuet. Referring to the sarabande, it reads: ‘restiren noch 8 Variationen nebst von 2 Violin und Bass’. This must have been copied from Source Y since Walther would have known it to be inaccurate had he knowledge of the original imprint.[16]

An interesting feature of Walther’s copy is its additional ornaments. Although there are a number of extra chutes and pincés, he also added written out embellishments in the form of ‘nottes perduës’ (to use Le Roux’s term) and the spatial arrangement of their containing bars and the manner in which they are accommodated suggests they were not present in his source but added after Walther completed his copy. His intention was possibly to notate the essence of spontaneity and fluidity that is most often associated with le goût. Quintessentially French, it was an aesthetic that was probably as difficult to assimilate then as it is today and such additions might also be seen as be evidential of the disparities that existed between two discrete musical languages. They also provide examples of how sources were sometimes altered to facilitate performance.

An intriguing aspect of the suite’s ornamentation is the inclusion of a number of extra tremblements appuyés that were notated differently to the few examples that are found in the 1705 volume. They consist of chute and tremblement symbols appended to the same note (e.g. bar 1 of Allemande), a combination that is not to be found in other French sources.[17] Similar examples are to be found in Germany, however, such as the Explication in Wilhelm Friedemann Bach’s notebook. These might have been Walther’s additions, although elsewhere he used appoggiaturas to notate the same effect (e.g. bar 11 of Courante). Nevertheless, their use suggests that Source Y was of German provenance. Another notational feature, however, indicates that it was not copied directly from the 1705 edition but from a secondary source that was very likely to have been French since it contains a system of shorthand that replaces the inner notes of chords with open noteheads to which are applied short curved stems. Although something similar had been used by Frescobaldi as far back as 1615, whereby curved stems allowed differentiation between voices in homophonic music, the notation used in P 801 appears to be exclusively French and was a simple yet elegant device that kept scores free from clutter.[18] There are no doubts that it was known to both Walther and the copyist of Source Y since neither discarded it in favour of more conventional notation, yet the general absence of other German examples indicates that it was neither preferred nor one that was in common use outside France. We can be certain, therefore, that P 801 was a copy of a source that was probably German, which in turn had been prepared from a French copy, and since the predominant left hand clef is F3, its hypearchetype must have been the 1705 edition.

The arrangement of the suite is also intriguing, since the sarabande is abandoned after its first couplet (432) and followed by the gigue before being picked up again for its next three doubles. The length of the complete sarabande and its increasing complexities might explain why this is so. It would have required considerable effort to copy all twelve couplets, and since the variations become more virtuoso as the sarabande progresses, they might have been considered difficult to such a degree that their exclusion occurred. The three extra couplets were therefore possibly an afterthought: they are not particularly difficult and at least provide a few of the more manageable and attractive variations.

P 801 might be dismissed as an inconsequential source of French harpsichord music, since the focus of interest for scholars lies in its J S Bach content. Yet the variant readings of Suite VII make it an important source of information on the dissemination of a specific musical language and provides an insight into an area of performance practice that today’s research has yet to address fully. It also indicates that both print and manuscript copies of Le Roux’s works were circulated as far as Germany. While this cannot be taken as an indication of Le Roux’s sphere of influence, it does suggest that his music was known outside his immediate environment in an edition that was afforded neither the vast distribution network nor advertising resources that Roger had at his disposal.


[1] Musica Repartita of Utrecht (1985) and Éditions Minkoff (1995) reproduced a cleaned-up facsimile of the Bibliothèque nationale; a similarly-treated version of the Washington copy is available from Broude Bros., New York; no facsimile exists of the Cambridge Amsterdam copy.
[2] Le Roux, ed. Fuller (1959), xxi.
[3] Tilney (1985).
[4] For example, the second Allemande (V) uses heavy brisure.
[5] Tuffery (2016), 28.
[6] D-B Mus. ms. Bach P 801.
[7] Schmidt (2000), 120.
[8] Ibid., 123.
[9] Rasch (2012), 5. Le Roux’s pièces are found in: 1708, 18; 1712, 43; 1716, 346; 1725, 69; 1735, 67; 1744, 48. Le Cène, Roger’s son-in-law and took over the business in 1723, adding approximately 100 further titles to the catalogue, including works by Locatelli and Tartini.
[10] Fuller (1959), xxi; McKean (2011b), 16.
[11] For example, Roger rescored the closing of the two-harpsichord gigue to provide a less confusing structure than that of the Paris edition.
[12] The date provided by the Staatsbibliothek, Berlin is 1710, but Beißwenger (1988, 22) suggests 1714. The content copied by Walter consists of Suites I and III, omitting both preludes, and the sarabandes and last minuet from Suite III. In addition, the first harpsichord part of Gigue is to be found.
[13] Both suites reveal over 150 omissions, including ornaments, ties and slurs and notes, and over 20 instances where an ornament has been misread or altered.
[14] Menuet (I, bar 17) contains parallel octaves between the soprano and bass, which he addresses by rewriting the first two beats, and Courante (I, bar 13) contains a rescored left hand to make an unusually large interval more manageable.
[15] ‘Suite di Clerambouloux’. Its contents consist of the first five movements of Suite VII, couplets 1-4 of the sarabande and the minuet. McKean (2011a), 8. A copy of this suite is reproduced in Appendix 2 (99).
[16] P 801, 442: ‘remaining still [are] 8 variations together with 2 violins and bass’.
[17] Brunold (1965, 10), records either the left descender or a petite note slurred to a trill. The combination is not found in Saint Lambert.
[18] For example, Frescobaldi (1615) Toccata Prima. F-Pn Néerlandais 58 contains works for carillon that were written by a Fleming in Paris during the 1620s or 1630s, and similar examples may also be found in the works of Chambonnièrs (1670, Allemande la Rare), Lebègue (1676, 1), and Dandrieu (1705, 1).