The author wishes to thank the editors, readers and many others who provided feedback on earlier versions of this article for their invaluable help and many suggestions.Google Scholar

1 Voltaire, Candide, ed. Sylviane Léoni (Paris, 1995), 144. ‘While waiting for dinner, Pococurante had a concerto performed. Candide found the music enchanting. “This noise”, said Pococurante, “can give half an hour's amusement; but if it lasts any longer it bores everyone, though no one dares to admit it. Music today is nothing more than the art of performing difficult pieces, and what is merely difficult gives no lasting pleasure‘” (Voltaire, Candide: Or Optimism, trans. John Butt, London, 1947, 119).Google Scholar

2 The earliest of Satie's letters to mention Socrate specifically ('Je m'occupe de la “Vie de Socrate”') dates from 6 January 1917; the last, in which he writes ‘Je remets l'orchestre au net’, is dated 10 October 1918. See Satie, Erik, Correspondance presque complète, ed. Ornella Volta (Paris, c. 2000), 273, 341.Google Scholar

3 ‘J'ai une frousse de “rater” cette oeuvre que je voudrais blanche & pure comme l'Antique’ (ibid., 273).Google Scholar

4 ‘Je travaille à la “Vie de Socrate”. J'ai trouvé une belle traduction: celle de Victor Cousin. Platon est un collaborateur parfait, très doux & jamais importun. Un rêve, quoi! […] Je nage dans la félicité. Enfin! je suis libre, libre comme l'air, comme l'eau, comme la brebis sauvage. Vive Platon! Vive Victor Cousin! Je suis libre! très libre! Quel bonheur!‘ (ibid., 277–8).Google Scholar

5 ‘C'est un retour vers la simplicité classique, avec sensibilité moderne. Je dois ce retour – aux bons usages – à mes amis “cubistes”. Qu'ils soient bénis!‘ (ibid., 325; English translation from Ornella Volta, Satie Seen through his Letters, trans. Michael Bullock, London and New York, 1989, 152).Google Scholar

6 ‘Je suis très content de mon travail’ (Satie, Correspondance, ed. Volta, 329).Google Scholar

7 ‘[…] mon oeuvre maîtresse’ (ibid., 347).Google Scholar

8 In the frontispiece of the first edition, the piece is generically defined as a ‘drame symphonique avec voix’. The French substantive ‘voix’ is indeclinable, so remains ambiguous about the required number of singers: one or more? The characters are four, but they never speak at the same time, so a single singer can perform the whole piece. Satie allowed performances with only one voice, but preferred the four sopranos, where each character corresponds to a different singer. On 16 April 1919 he explains: ‘“Socrate” est écrit pour soprani. Je désire qu'en “public” il soit donné comme je l'ai écrit… “avec 4 soprani”’ ('Socrate is written for sopranos. I wish it to be performed in public as I wrote it…with 4 sopranos'; Satie, Correspondance, ed. Volta, 361).Google Scholar

9 '“Socrate” – Drame symphonique avec 4 voix, … écrit sur les Dialogues de Platon …Google Scholar

… La traduction est de Victor Cousin. Cet ouvrage comporte 3 parties: …Google Scholar

… La première nous donne un portrait de Socrate par Alcibiade, portrait tiré du Banquet; …Google Scholar

… dans la deuxième partie, nous assistons à une promenade de Socrate & de Phèdre le long des bords de l'Ilissus. Cette partie est tirée de Phèdre; …Google Scholar

… la. troisième partie nous reconstitue la mort de Socrate, tirée de Phédon, & racontée par celuici …Google Scholar

… En écrivant cette oeuvre, … je n'ai nullement voulu ajouter à la beauté des Dialogues de Platon: … ce n'est, ici, qu'un acte de piété, qu'une rêverie d'artiste, … qu'un humble hommage …Google Scholar

… L'esthétique de cet ouvrage se voue à la clarté; … la simplicité l'accompagne, la dirige … C'est tout: … je n'ai pas désiré autre chose …' Paris, Bibliothèque Nationale de France (fonds Jane Bathori), unsigned and undated autograph manuscript, reproduced in Ornella Volta, Erik Satie (Paris, 1997), 139. See also Erik Satie, Quaderni di un mammifero, ed. Ornella Volta (Milan, 1980), 230.Google Scholar

10 Here, in the interests of brevity and convenience, I shall use the French title of the Platonic dialogue each one is taken from: Le banquet, Phèdre and Phédon.Google Scholar

11 In the letter of 1 December 1918 cited above, he also states that the four sopranos are to be ‘deux aigus, deux mezzo’ (Satie, Correspondance, ed. Volta, 347).Google Scholar

12 ‘[…] une seule oeuvre comme le Socrate se révèle à l'analyse aussi riche, sinon plus riche, en idées musicales qu'un monument comme la Tétralogie’. Robert Caby, ‘Erik Satie à sa vraie place’, La revue musicale, 214 (June 1952), 29.Google Scholar

13 ‘Sulla marcata preferenza di Satie per i trittici […] sono state formulate diverse ipotesi. […] Georges-Jean Aubry, 1916, riporta che Satie voleva fare con questi suoi trittici “le due manches e la bella”’ ('Various hypotheses concerning Satie's marked preference for triptychs […] have been formulated. […] Georges-Jean Aubry, 1916, reports that Satie wanted his triptychs to be like “two rounds and the decider”'). Satie, Quaderni di un mammifero, ed. Volta, 193.Google Scholar

14 ‘La critique génétique reste muette devant des oeuvres sans manuscrits comme devant des manuscrits sans réécriture’ (‘Genetic criticism remains mute before works without manuscripts, as it does before manuscripts without any rewriting‘). Almuth Grésillon, Éléments de critique génétique: Lire les manuscrits modernes ([Paris], 1994), 30.Google Scholar

15 These bars correspond to the sentence pronounced by Socrates which closes Le banquet.Google Scholar

16 For detailed discussion of these points, see the extended study of which this article represents a summary: Pietro Dossena, ‘Labirinti compositivi tra avant-texte e texte: Sulla genesi del Socrate di Erik Satie’ (dissertation, University of Milan, 2005).Google Scholar

17 Postquam surrexit Dominus and Si ego Dominus start with E, D, E, G, (G), A, but continue with A, G, A, G, G, A, B, C. These melodies are in the Graduale triplex, 164, 166, and in the Liber usualis, 660, 661. The second also exists in a transposition to the first mode: Graduale triplex, 885.Google Scholar

18 I am grateful to Pieter Mannaerts for this information.Google Scholar

19 This melody is in the Graduale triplex, 198, and in the Liber usualis, 780.Google Scholar

20 This term is especially used in jazz (or, more precisely, jazz-rock): ‘The simplest definition of a slash chord is “a triad over a bass note”‘ (Mark Levine, The Jazz Theory Book, Petaluma, CA, 1995, 104). A slash chord is written as follows: D/C is a D major triad superimposed on a C functioning as bass (but I also permit the perfect fifth C–G). The last chord of i is thus Em/D. This kind of symbolic notation is simple and is able to represent the harmony of the variants we are currently examining.Google Scholar

21 'La matière (Idée) & la main d'oeuvre (écriture). La “main d'oeuvre” est souvent supérieure à la matière.Google Scholar

Avoir le sentiment harmonique c'est avoir le sentiment tonal.Google Scholar

L'examen sérieux d'une mélodie constituera toujours, pour l'élève, un excellent exercice harmonique.Google Scholar

Une mélodie n'a pas son harmonie, pas plus qu'un paysage n'a sa couleur. La situation harmonique d'une mélodie est infinie, car une mélodie est une expression dans l'Expression. N'oubliez pas que la mélodie est l'Idée, le contour; ainsi qu'elle est la forme & la matière d'une oeuvre. L'harmonie, elle, est un éclairage, une exposition de l'objet, son reflet' (trans. Robert Orledge, Satie the Composer, Cambridge, 1990, 68).Google Scholar

22 ‘Je dis d'abord qu'il [Socrate] ressemble tout à fait à ces Silènes qu'on voit exposés dans les ateliers des sculpteurs et que les artistes représentent avec une flûte ou des pipeaux à la main [the text of bars 29–45], et dans l'intérieur desquels quand on les ouvre, en séparant les deux pièces dont ils se composent, on trouve renfermées des statues de divinités [bars 46–59].‘Google Scholar

23 See Dom Clément Jacob, ‘Erik Satie et le chant grégorien’, La revue musicale, 214 (June 1952), 87–8.Google Scholar

24 Léon Guichard, ‘Erik Satie et la musique grégorienne’, La revue musicale, 169 (November 1936), 334–5.Google Scholar

25 Reported in Satie, Quaderni di un mammifero, ed. Volta, 231.Google Scholar

26 Ibid., 262–3.Google Scholar

27 Satie, Correspondance, ed. Volta, 270, 719.Google Scholar

28 Ibid., 317.Google Scholar

29 Ibid., 294.Google Scholar

30 Ibid., 296, in a letter to Charles-René, a composer and professor of harmony, teacher of the young Ravel.Google Scholar

31 The following reflections can also be applied to the other two key passages of Socrate that I have analysed elsewhere (see note 16): the start of Phèdre and the start of Phédon.Google Scholar

32 This statement refers to quite general considerations; with regard to Socrate, it concerns especially the variants of the instrumental accompaniment, rather than those of the vocal melody. ‘Abstract piano’ is used here with acknowledgement of the fact that Satie did not actually compose at the keyboard.Google Scholar

33 This is the interpretative category I prefer. Different typologies of homogeneity can exist, of course: for example, a homogeneity in excess, or a homogeneity in essentiality. In Socrate only the latter type is found.Google Scholar

34 We could hazard a relationship between this median level and a median gradation of simplicity: this way, homogeneity and simplicity would simply present two faces of the same coin.Google Scholar

35 Note the insistent alliteration here, which causes an amplifying effect. As Ornella Volta suggests (Erik Satie, Écrits, ed. Volta, [Paris], 1977, 296), this sentence about flatness may well have a contingent origin: ‘Cette réflexion prend place dans un carnet de notes pour Socrate, entre l'ébauche d'une lettre d'explications, adressée au critique Jean Poueigh, et un brouillon de l'Éloge des critiques, inspiré par le même monsieur. Il est donc possible d'imaginer qu'ES se soit inspiré du nom de son persécuteur pour cet éloge de la platitude, également. En effet, dans son ouvrage sur les musiciens français d'aujourd'hui, Jean Poueigh fait remonter l'origine de son nom au “vocabule pouey qui, dans certaines régions des Pyrénées, désigne toute montagne terminée par une plate-forme”’ ('This reflection is found on a sketchbook for Socrate, between the sketch of a letter of explanation addressed to the critic Jean Poueigh and a draft of the Éloge des critiques, inspired by the same man. It is thus possible to imagine that ES took inspiration from the name of his persecutor for this eulogy of flatness too. In fact, in his book on contemporary French musicians, Jean Poueigh individuates the origin of his surname in the “word pouey that, in certain regions of the Pyrenees, designates any mountain ending in a platform”').Google Scholar

36 This assumes the existence of a median level not only in the texte, but also in the avant-texte. In the avant-texte such a median level is variable, still settling down; whereas in the texte, it is steady, fixed.Google Scholar

37 This is a possible cause for the non-homogeneity of the first variants of Le banquet in comparison with the final immobile structure (the texte, or, metaphorically, the definitive puzzle).Google Scholar

38 Expressing the same concept in terms of simplicity, we could state that Satie proceeds in the direction of a median simplicity.Google Scholar

39 For further information about the use of tetrachords in Phédon, see Orledge, Satie the Composer, 133, 136–7, and the illuminating analysis by Oliver Vogel (‘Socrate von Erik Satie: Eine Identifikation’, dissertation, University of Berlin, 1994, XIV–XV), showing the transformations and recurrences of the initial motif.Google Scholar

40 Obviously, ripples can be individuated not only in some variants or in the f.v. of a key passage, but also in other passages – either of the avant-texte or of the texte.Google Scholar

41 In a letter to Paul Collaer of 16 May 1920, Satie writes: ‘En écrivant “Socrate”, je croyais composer une oeuvre simple, sans la moindre idée de combat’ (‘Writing Socrate I thought to compose a simple work, with no idea of struggle‘). Satie, Correspondance, ed. Volta, 406.Google Scholar

42 Orledge, Satie the Composer, 133.Google Scholar

43 In this regard, see Sala, Emilio, ‘Dalla Bohème all‘avant-garde. Ancora nel segno dei fumisti’, Erik Satie e la Parigi del suo tempo, ed. Gianmario Borio and Mauro Casadei Turroni Monti, Nuovi percorsi musicali, 4 (Lucca, 2001), 29–44. In this article he clarifies – among other things – the fumiste origin of what is called (after Jankélévitch) ‘conformisme ironique’.Google Scholar

44 ‘Le paradoxe c'est que, tout en faisant mine de se modeler à l'image du plus grand nombre et de s'interdire toute expression subjective, Erik Satie n'a fait qu'attirer les regards sur sa propre singularité.‘ Satie, Correspondance, ed. Volta, 10.Google Scholar

45 An interesting perspective on the subject can be found in Vladimir Jankélévitch, L'ironie ou la bonne conscience (Paris, 1950), where the author speaks of Socratic irony in terms that can be applied to Satie and to the aesthetic of Socrate: the litotes (a figure of speech employed by the Greek philosopher) belittles, alludes, understands, but its aim is positive; as a matter of fact, it invites reflection and stimulates knowledge. The same concept is expressed by Marc Bredel, Erik Satie (Paris, 1982), 147–56.Google Scholar

46 Satie, Quaderni di un mammifero, ed. Volta, 298.Google Scholar

a On p. 1, staves 3–4, right, an inscription from Satie: ‘Système de quartes ou quintes justes’.Google Scholar

b Second version of a.Google Scholar

c The note in square brackets has been erased by Satie.Google Scholar

d The parentheses to the different key signatures are Satie's.Google Scholar

e I indicate the following two fragments with letters of the Greek alphabet because they do not represent proper variants, but are related to other variants. The notes ‘I # ’ and ‘3 #’ in α are Satie's.Google Scholar

f The most plausible reading of this is that the g in the second chord is natural and the letters ‘M.’ and ‘m.’ refer to the chords in the upper stave (respectively D major and D minor).Google Scholar

g The notes ‘Un # ’ and ‘Trois # ’ are Satie's.Google Scholar

h The A in the last chord is evidently a mistake; it should be a B.Google Scholar

i Satie often writes across double pages in his notebooks, so some variants (like this one) can cross from the left- to the right-hand page. The numbering of the bars is mine and refers to the bars of the published text (f.v.). Stave 6 here is indecipherable, but may well be similar to the upper stave of m1, at least in the initial bars.Google Scholar

j This variant is almost completely indecipherable. At bars 48 and 50 the notes with no head simply indicate the rhythm (legible or intuitable), leaving the pitches indeterminate; where possible, I have tried to give an approximate idea of the pitches.Google Scholar

k In this notebook there are six staves per page, usually grouped as two or three systems.Google Scholar

l The notes in square brackets are Satie's, deleted in ink and replaced by other notes. In this case, I have not distinguished two variants q and q′, as the corrections are few and not particularly relevant.Google Scholar

m The notes are very uncertain, especially at bars 51 and 53–9. This variant and the next are to be read with an implied F# in the key signature. The slur in square brackets at bar 58 has been erased by Satie.Google Scholar

n At bar 57, left hand, the # to b in the tenor must be an oversight: the note is b, not b#. At bar 58, right hand, the c″s are to be read as c#″s. The mark ‘∗ B’ is by Satie and is associated with ‘∗ A’ in u (see below): the two parts A and B are linked together to form v and the final version.Google Scholar

o The comparison of the right hand in u with the upper stave of a′ reveals very interesting similarities: in correspondence to each change of note in the cantus firmus, the right hand in u plays fourths, analogous to those in a′. Are these resemblances sufficient to establish a connection between the two – which are distant in time, being respectively the third and the penultimate variant? In the synoptic scheme (see Figure 1) I have not underlined this similarity, but the reader should bear in mind that a more daring interpretation could establish a direct relationship.Google Scholar

p I consider as the final version the autograph manuscript (in the archives of Prince Louis de Polignac, but also existing in a facsimile reproduction by Eschig), jointly with the first printed edition of the vocal score (Éditions de La Sirène, 1920). The problematic aspects related to the selection of a definitive text (or texts?) of Socrate do not significantly affect my study of bars 46–59 of Le banquet.Google Scholar

q I do not include m1 in this group because it presents – in the upper stave – an alternative version of the left hand, diverging from the bare chordal sequence present in b, a′, c, l.Google Scholar

r In s′ this texture appears only in bars 55–6 and 59.Google Scholar