1 Sylvia Kahan, Music's Modern Muse: A Life of Winnaretta Singer, Princesse de Polignac (Rochester, NY, 2003), 57.

2 Gabriel Fauré: His Life through his Letters, ed. Jean-Michel Nectoux, trans. J. A. Underwood (London, 1984), 199–201. In 1888 Montesquiou provided the text for the ad libitum chorus to Fauré's Pavane, composed in the autumn of 1887 and dedicated to the Comtesse Greffulhe. His words are a spoof of Verlaine's Fêtes galantes, particularly Clair de lune and Mandoline (see ibid., 130–4).

3 Marguerite de Saint-Marceaux, Journal, 1894–1927, ed. Myriam Chimènes (Paris, 2007), 1288. Many salon performances of Wagner are documented in newspapers such as Le temps, viewable online through <http://gallica.bnf.fr>.

4 On Bagès's death, Breville wrote to Marguerite de Saint-Marceaux: ‘My poor dear friend Bagès died this morning […] The blow has been so sudden that I can't take it in. […] You know what a fine person he was. For thirty years he was my absolute friend, constant and devoted. My grief is profound’ (Saint-Marceaux, Journal, ed. Chimènes, 525n. (7 November 1908): ‘Mon pauvre cher ami Bagès est mort ce matin […] Le coup a été si subit que je n'y puis croire. […] Vous savez comme il était bon. Depuis trente ans il était pour moi l'ami véritable, constant et dévoué. Mon chagrin est profond’). Unless otherwise acknowledged, all translations are by the present authors.

5 Michel Duchesneau, L'avant-garde musicale et ses sociétés à Paris de 1871 à 1939 (Liège, 1997), 248.

6 Pierre de Bréville, ‘Quelques souvenirs’, La revue musicale, 201 (September 1946), 229: ‘Il habitait alors à Marnes, et souvent quelques-uns de nous se retrouvaient chez lui. Parmi eux parfois Maurice Bagès artiste né qui, par modestie, ne voulut jamais accepter que le titre d'amateur bien que, cédant aux sollicitations de Chevillard, il dût un jour interpréter le rôle de Siegfried au concert Lamoureux. C'est dans une de ces réunions amicales que Fauré eut l'occasion de l'entendre chanter Nell et Après un rêve. Le timbre expressif de sa voix, son incomparable diction, sa musicalité l'enchantèrent. “Chantez-vous d'autres mélodies de moi? lui demanda-t-il. Je les aime et chante toutes” lui répondit Maurice Bagès, et de ce jour Fauré l'adopta comme son meilleur interprète, lui confia à la Société nationale la création de Clair de lune et Au cimetière, et lui demanda de l'accompagner à Londres lors de son premier voyage en Angleterre pour y faire connaître ses oeuvres.’ The orthography of Breville's name raises a query: though it was invariably printed as ‘Bréville’, the accent is absent from his signature on numerous autograph manuscripts in the Conservatoire Royal de Bruxelles and the Bibliothèque Nationale de France, except for just one (in Brussels) which curiously shows the accent written then deleted. Mimi Segal Daitz has kindly corroborated this trait (the accent absent) from his signatures on letters and other manuscripts (email to Roy Howat of 13 March 2013).

7 Jean-Michel Nectoux, Gabriel Fauré: A Musical Life, trans. Roger Nichols (Cambridge, 1990), 537, 540–4.

8 Georges Servières, Gabriel Fauré (Paris, 1930), 26: ‘Mais c'est le ténorino mondain Maurice Bagès qui soupira le Clair de lune et La fée aux chansons […] et déclama les strophes de Jean Richepin: Au cimetière.’

9 Michel Faure and Vincent Vivès, Histoire et poétique de la mélodie française (Paris, 2000); Helen Abbott, Between Baudelaire and Mallarmé: Voice, Conversation and Music (Aldershot, 2009); Katherine Bergeron, Voice Lessons: French Mélodie in the Belle Epoque (New York, 2010).

10 Correspondance de Marcel Proust, ed. Philip Kolb, 21 vols. (Paris, 1970–93), i, 338; cited in Gabriel Fauré: Correspondance, ed. Jean-Michel Nectoux (Paris, 1980), 207: ‘Bagès m'avait laissé tout son Fauré. Mais la moitié est en manuscrits de Fauré, le reste imprimé, mais porte des dédicaces, et j'ai eu si peur de les abîmer que j'ai tout renvoyé.’ Most of these scores are now untraced, except for a handful of professional hand copies and early prints now in the library of the Conservatoire Royal de Bruxelles, and an autograph of C'est l'extase (see note ³² below).

11 Gabriel Fauré, Complete Songs, i: 1861–1882, ed. Roy Howat and Emily Kilpatrick (London and Leipzig, 2014); vols. ii–iv are in press or preparation.

12 Gabriel Fauré, Mélodies et duos, i: Premières mélodies, 1861–1875, ed. Jean-Michel Nectoux and Mimi Daitz (Paris, 2010), 150.

13 Roy Howat and Emily Kilpatrick, ‘Editorial Challenges in the Early Songs of Gabriel Fauré’, Notes, 68 (2011), 239–83 (p. 271). An alternative rationale for keys and transpositions is accordingly set out and followed in Fauré, Complete Songs, i, ed. Howat and Kilpatrick.

14 Exceptions include Ravel's Sainte and the Debussy anthology Douze chants (high- and low-voice editions), plus a high-voice transposition of his second set of Fêtes galantes, all published by Durand in 1906–7. The anthology includes Le jet d'eau from Cinq poëmes de Baudelaire, a cycle that Debussy, in a cordial letter of 1898, had already unhesitatingly authorized one E. Schmitt (presumably a baritone or bass) to transpose as necessary: see Claude Debussy: Correspondance, 1872–1918, ed. François Lesure and Denis Herlin (Paris, 2005), 430. Several of Debussy's earlier songs were also printed singly in several keys. Although his Ariettes oubliées were printed only in original keys, the Debussy Nachlass included a non-autograph manuscript of the second one, Il pleure dans mon coeur, transposed down a tone to F♯ (now in the Musée Debussy at Saint Germain-en-Laye), as well as one autograph of his early 1880s setting of En sourdine a tone lower than his other three autographs of that same setting (which remained unpublished during his lifetime); see Marie Rolf, ‘Debussy's Settings of Verlaine's “En sourdine”’, Perspectives on Music: Essays on Collections at the Humanities Research Center, ed. Dave Oliphant and Thomas Zigal (Austin, TX, 1985), 205–33 (p. 215).

15 The 1894 high-voice edition printed the second and third songs in F minor and F major instead of their manuscript keys of F♯ minor and G♭ major.

16 Exceptionally, Hymne appears in what became the high-voice key of G; in 1890 Hamelle transferred Hymne in G to the high-voice edition of the first collection, replacing it with the medium-voice key of F.

17 Beyond the scope of the present article, the labelling and significance of ‘original key’ in song publishing, and its later, more academically charged, association with historically informed performance practice and critical editing, constitute a fascinating topic that invites more study.

18 Bibliothèque Nationale de France, Département de Musique (henceforth BnF Mus.), MS 17777.

19 See Howat and Kilpatrick, ‘Editorial Challenges’, 273 and 273n., as well as Saint-Marceaux, Journal, ed. Chimènes (passim).

20 Among the many early songs for which more than one authorial key is known, autograph sources in different keys survive for Puisque j'ai mis ma lèvre, Au bord de l'eau and Les matelots; manuscripts also survive in different keys from that (or those) of original publication for Le papillon et la fleur, Mai, La rançon, La fée aux chansons and Poème d'un jour (see note ¹⁵ above). See also note ³⁴ below concerning La bonne chanson.

21 Nectoux and Daitz perhaps overlooked these scribal copies in arguing that ‘Fauré had had to accept this custom with regard to transpositions, apparently seemingly [sic] deciding to prepare them himself’ (Fauré, Mélodies et duos, i, ed. Nectoux and Daitz, 159). A personal communication kindly sent by Jean-Michel Nectoux amplifies this (email to Roy Howat, 24 January 2011): ‘Dans les archives Hamelle, on lit plusieurs fois pour les tons transposés “ms pas faits”, parfois, Fauré a fait attendre ces transpositions à son éditeur plusieurs années, Hamelle conservant le n° de cotage original, bien que la gravure soit très postérieure’ (‘In the Hamelle archives one often sees noted for the transposed keys “ms. not made”[:] sometimes Fauré made his publisher wait several years for these transpositions, Hamelle using the original plate number even though the engraving was much later’).

22 Fauré, Mélodies et duos, i, ed. Nectoux and Daitz, 150.

23 Gabriel Fauré: His Life through his Letters, ed. Nectoux, 275–6. Any argument that Fauré's fame after 1905 ‘enabled him to impose his point of view’ (whatever it was) also circumvents the lurking issue that by the 1880s Fauré's reputation was well established and spreading internationally, whereas his late songs, in a more elliptic idiom, never sold as well – a point cynically rubbed home by Hamelle's reported retort in those later years to Roger-Ducasse, who had asked him if he didn't regret losing Fauré's custom: ‘J'm'en fous, j'ai c'qui s'vend’ (‘Couldn't give a toss, I've got what sells’; Jean-Michel Nectoux, Gabriel Fauré: Les voix du clair-obscur (Paris, 1990), 280).

24 In the teeth of Fauré's documented wishes, Hamelle reprinted the pieces with appended titles, and detached the last of them – which Fauré had given him for free – to market separately as the ‘Huitième Nocturne’; see the Preface to the Peters Edition of the Pièces brèves, ed. Roy Howat (Peters Edition, London, EP 7601, 2003). After 1904, Fauré continued dealing with Hamelle over reprints or re-editions of works that Hamelle already owned. His fiery letter about the A major Violin Sonata is in BnF Mus., NLA 12 (222).

25 Nectoux and Daitz acknowledge this, noting the problems inherent in ascertaining ‘original’ keys of the early songs, as ‘some manuscripts might have been transposed [by Fauré] for use by this or that friend or performer’ (Fauré, Mélodies et duos, i, ed. Nectoux and Daitz, 158) – a statement that immediately undermines any in-principle assertion against transposition.

26 BnF Mus., MS 17787 (1).

27 Now in the Médiathèque Mahler, Paris, ex coll. Alfred Cortot; the title page is annotated, probably in Fauré's hand, ‘pour ténor ou soprano’. The other copy, in a different hand, served for many years as Hamelle's hire copy (now in BnF Mus., Vma MS 1156).

28 Hand copy now in the library of the Conservatoire Royal de Bruxelles, shelf mark 43.928; its title page bears a later annotation by Pierre de Breville: ‘C'est sur cette copie que fut chanté pour la première fois le “Clair de lune” à la Société Nationale par Maurice Bagès.’ This must refer to the orchestral première, as no other early Société Nationale performance is on record, nor is any first performance with piano traced. Moreover, the B♭ minor manuscript copy was clearly a pair with the orchestral conducting manuscript, despite the different keys: its title page exactly matches the latter (though the music is in a different hand), and it shows the same rehearsal cue letters in the same coloured pencil and hand, together with pencilled instrumental indications (partly erased) exactly matching those in the orchestral score. This would have been a rare instance of Bagès transposing from copy (the first edition not yet being in print), for he appears thereafter to have been always provided with high-voice copy, sometimes made for the occasion (perhaps the fruit of experience, Clair de lune having been his first important concert collaboration with Fauré).

29 The three early songs that stand out as prefiguring this trend are Fauré's Baudelaire settings of the 1870s.

30 Gabriel Fauré: His Life through his Letters, ed. Nectoux, 186. In face of many singers’ regret that Fauré did not set more Verlaine than he did, we might be grateful that he persisted at all. (In the event, he set exactly the same total of Verlaine poems – seventeen – as Debussy did, six of them the same poems.)

31 Kahan, Music's Modern Muse, 62. Kahan also traces the fascinating connections between Fauré, Robert de Montesquiou, Maurice Bagès and the Prince and Princesse de Polignac (pp. 69–74), and notes the small but significant role played by the ‘Venetian’ songs in bringing about the Polignacs’ marriage.

32 This manuscript was originally given to the princesse: Fauré explained that ‘the manuscript that I am sending you […] has been through the hands of the engraver. I shall do another one for you, and I am only sending you this one in order to acquaint you with the last song as soon as possible’ (Gabriel Fauré: Correspondance, ed. Nectoux, 187: ‘le manuscript que je vous envoie […] a passé par les mains du graveur. J'en ferai un autre pour vous et je ne vous fais parvenir celui-ci que pour vous faire connaître plus tôt la dernière mélodie’). If Fauré did write out another copy it has not been traced; the princesse must subsequently have presented the engraving manuscript to Bagès. Acquired later by Alfred Cortot, it is now in the Parisian collection of Eric Van Lauwe.

33 This manuscript was originally given to the princesse: Fauré explained that ‘the manuscript that I am sending you […] has been through the hands of the engraver. I shall do another one for you, and I am only sending you this one in order to acquaint you with the last song as soon as possible’ (Gabriel Fauré: Correspondance, ed. Nectoux, 187: ‘le manuscript que je vous envoie […] a passé par les mains du graveur. J'en ferai un autre pour vous et je ne vous fais parvenir celui-ci que pour vous faire connaître plus tôt la dernière mélodie’). If Fauré did write out another copy it has not been traced; the princesse must subsequently have presented the engraving manuscript to Bagès. Acquired later by Alfred Cortot, it is now in the Parisian collection of Eric Van Lauwe.

34 Even in La bonne chanson – which is a cycle – Fauré seems not to have placed particular importance on the original key sequence. The 1908 medium-voice edition variously transposes songs 2–9 between a semitone and a minor third down (leaving no. 1 untransposed), while Fauré also copied out the third and seventh of the songs a semitone down (probably in the late 1890s and presumably for the benefit of a particular singer; these manuscripts are now in the Sibley Music Library, Eastman School of Music, University of Rochester, NY). Meanwhile, Fauré's 1898 transcription for chamber ensemble transposes just the first two songs up a tone; Bagès (who premièred this version) was provided with scribal transpositions of these (now in the library of the Conservatoire Royal de Bruxelles, shelf marks E 43.925 and 931).

35 That of Green also served as engraving copy for the original high-voice edition séparée.

36 Moreover, Hamelle, who seems to have shared with other publishers a phobia of ‘dense’ key signatures, would never have opted for F♯ major on his own account (besides which, it would have used more ink). Fauré's Baudelaire setting Chant d'automne, originally published (by Choudens) in D♭ minor (with its coda in D♭ major) appears in the Hamelle collections in C minor/major; on this matter see Howat and Kilpatrick, ‘Editorial Challenges’, 270.

37 His concerns are set out in the long letter to Julien Hamelle of 1907 referenced in note ²³ above.

38 This revealed itself particularly in rehearsal and concert performance with the baritone Kurt Ollmann (song recital with Roy Howat at The Forge, London, on 20 June 2012, songs by Fauré, Debussy, Chabrier and Ravel).

39 Graham Johnson explores the connections between Montesquiou, Villiers and Fauré: see his Gabriel Fauré: The Songs and their Poets (Aldershot, 2009), 165.

40 Fauré's only other song to specify voice-type is the cantique Noël, op. 43 no. 1 (originally for tenor or soprano), which shares its opus number with Nocturne, an odd pairing for which no explanation is known. For Nocturne it seems likely that Fauré had a specific voice in mind; although there is nothing to link the song to a particular individual (no première is documented), one possible candidate is the fine contralto Julie Lalo (wife of the composer Édouard), who premièred Fauré's Lamento and to whom he dedicated Les matelots and Tristesse.

41 Graham Johnson's remark that the altered piano tessitura in the C major transposition was necessary ‘to stop it falling off the end of the keyboard’ (Johnson, Gabriel Fauré, 164), perhaps made in jest, overlooks the fact that the A♭ version shows the same adjustment; in any case, even with the piano at the higher octave both these versions would still lie within the compass of any post-1850 piano.

42 See, for example, BnF Mus., Vmg 22543 (2).

43 The manuscript appears to be a companion to one of Les présents (formerly in the collection of the Bibliothèque Musicale François-Lang, Abbaye de Royaumont): signed and dated 27 January 1892 and very similar in presentation, the latter similarly shows Hamelle's house stamp and no engraver's marks. Les présents, however, was indeed issued in that key (E♭) as an édition séparée before 1897; why no equivalent édition séparée appeared from the Nocturne manuscript remains a mystery.

44 Meanwhile, a feuilleton publication in E♭ had appeared in Les annales politiques et littéraires, 1173 (17 December 1905). Oddly, the last edition of the song during Fauré's lifetime was again in A♭ (clearly taken from the 1897 edition), printed in the February 1909 Album de Musica (no. 77), some months after the reissued Hamelle Collections had replaced this key with E♭. The most likely explanation is that the Album was prepared before the amendments were made for the Hamelle reissue.

45 Public workshop on 8 February 2013 at the Royal Academy of Music, London, presented by the present authors with postgraduate singers including Bradley Smith (a tenor with a vocal range and flexibility akin to that of Bagès) performing and exploring the source and key history of songs including Nocturne and La rose.

46 Jean-Michel Nectoux, ‘Deux interprètes de Fauré: Émilie et Édouard Risler’, Études fauréennes, 18 (1981), 3–25 (p. 11): ‘Nocturne on ne lui chante jamais et il l'aime, je le travaillerai mais il tâchera de m'avoir le ton original qui est très bas, car il l'aime en contralto.’

47 Jean-Michel Nectoux, ‘Deux interprètes de Fauré: Émilie et Édouard Risler’, Études fauréennes, 10; regarding Accompagnement, see Études fauréennes., 10, 12–15.

48 Gabriel Fauré: His Life through his Letters, ed. Nectoux, 142.

49 Johnson, Gabriel Fauré, 170.

50 Bagès's own copy of Au cimetière, a scribal copy (along with one of Spleen) evidently made from the autograph engraving manuscript, is now in the collection of the Conservatoire Royal de Bruxelles (shelf mark 43.927); like that of Clair de lune, it is annotated by Breville as having served for the première. The copy of Spleen (shelf mark 43.926) is at original medium-voice pitch (evidently copied from Fauré's autograph before the latter had been fully marked up for the engraver), perhaps an advance copy for perusal, or even just a keepsake.

51 In the letter cited in note ¹⁰ above, Proust called the song ‘truly awful’ (‘vraiment affreux’); this extract is quoted in Gabriel Fauré: His Life through his Letters, ed. Nectoux, 215 (Gabriel Fauré: Correspondance, ed. Nectoux, 212).

52 With characteristic waywardness, the f for voice at bar 13 appears only in the second collection's medium-voice edition, which places the ensuing f for voice at bar 17, whereas the high-voice edition prints it at the last note of bar 16. A later reprint of the original édition séparée (post-1908) incorporates most of the second collection's amendments (notably to articulation and pedalling), but then introduces a new reading of dynamics over bars 16–17 (decrescendo to p) and retains the original edition's dynamics over bars 28–31 (as in Example 4a). This aptly demonstrates the piecemeal nature of Fauré's revisions, and the associated headaches for the critical editor.

53 A related instance of amended dynamic shape can be observed in the earlier Aurore, op. 39 no. 1 (1884). A feuilleton publication in the Album du Gaulois, which appeared on 1 February 1885, clearly derived either from the composer's manuscript or an early proof, shows a crescendo to f at bar 31, with the reprise of the opening material at bar 32 then indicated mf for voice and p for piano (all other printed sources show a decrescendo preceding the key change at bar 31, with the reprise then indicated f for both). The same source reciprocally gives the dynamic at bar 42 as f, preceded by a crescendo; later sources instead show a decrescendo to p. A manuscript of En sourdine similarly shows the song's final reprise matching the dynamic marking of bar 1 (p); printed sources progressively amend this to the f of the 1908 edition, in a crucial rethinking of the song's dramatic structure (manuscript copy in private collection).

54 Unpublished letter, undated, BnF Mus., shelf mark NLA 264: ‘Mon cher ami, tu as très bien fait de me donner ton avis: seulement je crois que si nous avions vu ensemble la nouvelle fin je te l'aurais peut-être fait comprendre autrement. D'abord (il y a peut-être aussi une erreur de copie) les mots: “avec la beauté” doivent se chanter en diminuant […] puis l'accompagnement est d'une sonorité pleine mais douce; il faut de la pédale pour bien soutenir l'arpège de la basse et n'accentuer que ceci: […] les deux fois. La première fin ne me satisfaisait pas: si j’étais arrivé à Cuy deux jours plus tard je n'aurais pas apporté la mélodie telle que tu l'as vue là-bas. Cela me paraissait écourté et tu dois te souvenir que j'avais prolongé d'une mesure la période salua la Fleur. Mais ça ne me suffisait pas! Je suis resté tracassé jusqu’à ce que j'ai pu [overwriting “j'ai”] fait [faire] le changement complet. Je crois que dans quelque temps, quand tu auras oublié la première version tu seras de mon avis. Je dois dire que S^t Saëns ne connaît que la seconde et qu'il n'a pas paru choqué: ce qui en fait supposer que si tu n'avais pas d'abord chanté la première fin la seconde ne t'aurait pas causé de surprise. Tout ça n'empêche pas que je me trompe peut-être aussi: mais ce qui me rassure c'est ce tracassin qui ne m'a quitté qu'après que j'ai en changé ma péroraison.’

55 Jean-Michel Nectoux has discussed this characteristic, noting how many of Fauré's manuscripts include bar counts at the ends of pages, sections or systems (Gabriel Fauré: A Musical Life, trans. Nichols, 488–9). See also the critical commentary to Peters Edition EP 71904 of Fauré's Barcarolles (ed. Howat, London, 2011).

56 It is not clear whether the surviving autograph manuscript (used for engraving the first edition) might reveal any remnants of the original ending. Sold at Sotheby's on 12 April 2007 (Lot 39), its first music page (bars 1–9) is viewable through <www.sothebys.com>. In response to a query sent via Sotheby's, the buyer declared himself presently unable to allow access to it, having ‘already granted access to another party’ (email of 3 February 2010 conveyed to Roy Howat via Sotheby's).

57 The op. 51 songs had first appeared in the 1897 second collection (which comprised 25 songs) before being shifted to the third collection in 1908; see Nectoux, Gabriel Fauré: A Musical Life, trans. Nichols, 531n., and Fauré, Mélodies et duos, i, ed. Nectoux and Daitz, 163.

58 Fauré's manuscripts of many works bear witness to struggles with endings, and the pianist Marguerite Long recalled Debussy wryly remarking of Fauré, ‘Il ne sait pas finir’; see her Au piano avec Gabriel Fauré (Paris, 1963), 105. The forthcoming Peters critical edition addresses analogous structural problems in the piano postlude ending La fée aux chansons.

59 Gabriel Fauré: His Life through his Letters, ed. Nectoux, 252.

60 One of Fauré's first and most important reforms as the director of the Conservatoire was a complete overhaul of the teaching of singing. Importantly, he raised art song to a central place in the curriculum for the first time.

61 This sort of reactive revision is not atypical of Fauré, being also manifest in his piano and chamber music. See Roy Howat's critical editions of Fauré's Nocturnes and Barcarolles (Peters Edition, London, EP 7659 and 71904, 2006 and 2011) and of his First Piano Quintet (Editions Hamelle, Paris, HA 9690, 2006).

62 Through confusion between Fauré's London and Paris publishers Metzler and Hamelle, Pleurs d'or appeared from Metzler with the opus number 71 just as Hamelle published it as op. 72, having reserved 71 for the Thème et variations. Both publishers accordingly renumbered the Thème et variations as op. 73, though Hamelle's advertised work lists continued for many years thereafter to list it as op. 71.

63 Those in Mandoline and Arpège appear in all known sources; those in Pleurs d'or appear in only the Hamelle editions, being absent from the manuscript and the Metzler edition.

64 Debussy accompanying Mary Garden, Gramophone & Typewriter Co., recorded in May 1904: L'ombre des arbres (Ariettes oubliées), G&T 3077F-11, matrix no. 33450; various reissues on CD include EMI and Pierian, and most recently Marston Records.

65 See note ³⁴ above. Puisque l'aube grandit, transposed to A for this version, would have been equally manageable by the string instruments in its original key of G; the opening Une sainte en son auréole does lie better for strings in its transposed key of B♭ than in its original A♭, and exploits the second violin's lowest notes – though it can be made playable in the original A♭ by switching some lines with the viola, as has been done (tacitly) in an anonymously edited 2011 Hamelle edition of this version of La bonne chanson transposed for medium voice.

66 The recent Peters complete Chopin edition deals with the question of complete and mutually incompatible versions (an endemic issue in Chopin's music) by selecting and editing a single principal source, while showing other viable versions complete in the Appendix (see in particular Christophe Grabowski's edition of the waltzes, EP 7575). The series editors, John Rink, Jean-Jacques Eigeldinger, Jim Samson and Grabowski, explain their overall editorial methodology thus (as printed in the commentary to each volume in the series): ‘The Complete Chopin is based on two key premises. First, there can be no definitive version of Chopin's works: variants form an integral part of the music. Second, a permissive conflation of readings from several sources – in effect producing a version of the music that never really existed – should be avoided.’ See also the final chapter (‘The Chopin “Problem”: Simultaneous Variants and Alternate Versions’) of Jeffrey Kallberg, Chopin at the Boundaries: Sex, History and Musical Genre (Cambridge, MA, 1998), 215–30.

67 See the brief explanation of editorial method at <www.schubert-ausgabe.de> (‘Project’). The new Peters Edition also presents two versions of Fauré's early song S'il est un charmant gazon (long published as Rêve d'amour; see Howat and Kilpatrick, ‘Editorial Challenges’, 255–6).