^* This is the revised version of a paper discussed at the MLA Marcel Proust Seminar, San Francisco, December 1975. I should like to express my thanks to Marcel Muller and Michael Riffaterre for their careful reading and criticism of the first draft.

¹ It would be a tedious exercise to attempt an exhaustive listing of works dealing with Proust's imagery. We may mention, among standard works, Stephen Ullman's The Image in the Modern French Novel (Oxford: Blackwell, 1960), and Victor Graham's The Imagery of Proust (Oxford: Blackwell, 1966). More recently, the best writing on this subject has been by Gérard Genette (“Proust palimpseste,” Figures, Paris: Seuil, 1966, and “Métonymie chez Proust,” Figures III, Paris: Seuil, 1972); and by Jean-Pierre Richard (Proust et le monde sensible, Paris: Seuil, 1974). For a discussion of the epistemological implications of metaphor and metonymy in Proust and others, see Paul de Man, “Semiology and Rhetoric,” Diacritics, 3, No. 3 (1973), 27-33.

² Pierre Fontanier, Les Figures du discours (Paris: Flammarion, 1968), p. 384 (my italics).

³ See Gérard Genette, “Métonymie chez Proust,” p. 58.

⁴ Leo Spitzer, “Le Style de Marcel Proust,” Etudes de style (Paris: Gallimard, 1970), p. 407. Subsequent page references to Spitzer's essay will be given in parentheses in the text.

⁵ A la recherche du temps perdu, édition de la Pléiade, 3 vols. (Paris: Gallimard, 1954), i, 39. Subsequent page references to the novel will be given in parentheses in the text, citing volume and page number of the Pléiade edition. Unless otherwise stated, all italics are my own. I should say a word about my referring to the protagonist of the novel as “Marcel.” This is in no way meant to suggest that I equate the protagonist with Proust himself; it is simply a convenient way of designating the protagonist, without having recourse, each time, to the longer and more awkward common noun. (I generally avoid the term “hero” because of its cultural connotations.) The use of “Marcel” has become standard practice in the work of several contemporary critics—notably Gérard Genette —who do not equate the protagonist with the author of the novel.

⁶ This paradoxical status of the parenthesis, as well as its extremely frequent occurrence in Proust's sentences, shows how different Proust's use of it is from the classical model. According to the rhetoricians (e.g., Fontanier), the parenthesis must be used with circumspection and sobriety, because it disperses the reader's attention and necessarily produces “l'embarras, l'obscurité, la confusion.” When it is used, it must be “courte, vive, rapide,” like the fleeting shadow of a bird in sunlight (Fontanier, pp. 385-86). Proust's parentheses are not only frequent but often very long, sometimes longer than the sentence in which they are inserted; furthermore, they may themselves contain parentheses, like those Russian dolls that enclose ever smaller replicas of themselves. See, for example, i, 59-60, where the description of the stained glass windows of the Combray church is interrupted by a parenthesis (“et dans le reflet oblique et bleu duquel, parfois …”), inside which is another parenthesis, set off by dashes. These two pages, incidentally, contain four separate parentheses (counting the double parenthesis as one), ranging in length from one to ten lines.

⁷ “Leo Spitzer et la lecture stylistique,” preface to Etudes de style, p. 32.

⁸ “Literature and Language,” Essays in Criticism, 7, No. 2 (1957), 129.

⁹ In Communications, 8 (1966), 6-15. An English translation of the article is to be found in New Literary History, 6, No. 2 (1975), 237-72.

¹⁰ For the distinction between singulative and iterative sequences, see Figures III, pp. 145-46.

¹¹ See his Proust et le roman (Paris: Gallimard, 1971), pp. 385-87. Tadié restricts his definition of the parenthesis in this context to that of a “récit dans le récit,” however, even though not all sequence-level parentheses are in fact “récits.” See also A. Ferré, “La Ponctuation de Marcel Proust,” Bulletin de la Société des Amis de Marcel Proust et de Combray, 7 (1957). Ferré remarks that for Proust, “La parenthèse est …plus qu'un signe graphique: quelque chose comme un symbole du mouvement de pensée et de composition qui lui est propre. Des sections entières de son œuvre peuvent être considérées comme de grandes parenthèses dans la progression de la narration” (p. 324; quoted by Tadié, p. 386).

¹² See Auerbach's brilliant analysis of Homeric flashbacks (and digressions in general) in “The Scar of Odysseus,” in Mimesis: The Representation of Reality in Western Literature (Princeton: Princeton Univ. Press, 1953), pp. 3-7.

¹³ “Quand il a une explication à donner, Balzac n'y met pas de façons; il écrit: ‘Voici pourquoi’; suit un chapitre” (Contre Sainte-Beuve, précéde de Pastiches et mélanges et suivi de Essais et articles, Paris: Gallimard, 1971, p. 271). His critique of Balzac notwithstanding, Proust himself was not above resorting to “voici pourquoi” even in the early, stylistically “exquisite” sections of A la recherche. Cf. i, 208, where the Balzacian formula precedes the first mention of “la petite phrase”; for a later variation, see ii, 818, where a narrative digression about Mme de Cambremer is introduced by the sentence: “Mais pour les cousins de Ch'nouville, voilà” (my italics). These are, however, only exceptions that prove the rule; with digressions as with other things, Proust prefers indirection to explicit designation.

¹⁴ Two other traditional kinds are the descriptive pause—which, as we mentioned above, Proust almost never used as such—and the portrait of a character, including at times a complete case history, introduced just as the character has made or is about to make his appearance on the scene. Proust, despite his celebrated perspectivism, which makes characters appear as they are perceived by a specific observer (see the famous portrait of Charlus at Balbec, i, 751-53), nevertheless presents quite a few traditional portraits in the middle of his sequences. See the presentations of M. de Bréauté (ii, 504), of the ambassadrice de Turquie (ii, 534), of M. and Mme de Vaugobert (ii, 642-46), etc.

¹⁵ Laurence Sterne, Tristram Shandy, Bk. i, Ch. xxii.

¹⁶ Here again, one finds counterexamples: occasionally, Proust emphasizes his digressions in a manner similar to Sterne's. See, e.g., ii, 651-52, where “Monsieur l'auteur” engages in a bantering dialogue with “Monsieur le lecteur” about his digressive tendencies; see also iii, 789, where the narrator calls attention to a “parenthesis” (actually a digression) he has just finished, and authorizes himself to begin another, “even longer” one. Negation and emphatic affirmation are, it is true, but two sides of a single coin.

¹⁷ The technical term proposed by Gérard Genette is “prolepse répétitive (Figures III, pp. 111-12). Genette's term for the flashback is ”analepse complétive“ (Figures III, pp. 92-93).

¹⁸ See S/Z (Paris: Seuil, 1971), p. 39; see also Figures III, p. 114.

¹⁹ A longer, comparative study of sequence-level parentheses might allow us to categorize various kinds of texts, ranging from those that contain almost no parenthetical material to those, like Proust's, whose distinguishing feature is the presence of frequent, long sequence-level parentheses with a high degree of parentheticalness. In such texts, we might find that the very distinction between sequence and parenthesis breaks down, or, more exactly, that the overriding presence of parentheses makes the notion of sequence itself problematical. This kind of breakdown might be seen as paralleling what Genette has diagnosed as the breakdown, in modern texts, of the distinction between récit and discours. (See “Frontières du récit,” Figures II, Paris: Seuil, 1969.)

²⁰ For Muller's discussion of roman abstraits, in a context different from ours, see Les Voix narratives dans “A la recherche du temps perdu” (Geneva: Droz, 1965), pp. 66-77.

²¹ For an analysis of the functioning of exempla and didactic narratives in general, see Susan Suleiman, “Le Récit exemplaire,” Poétique, forthcoming in 1977.

²² “Semiology and Rhetoric,” Diacritics 3, No. 3, 32.

²³ Proust et le monde sensible (Paris: Seuil, 1974), p. 164 (my italics).

²⁴ See the chapter entitled “Anti-Logos,” in Proust et les signes, 2^e éd. augmentée (Paris: PUF, 1970). A similar argument, conducted along different lines, can be found in G. Genette, “Proust palimpseste,” Figures (Paris: Seuil, 1966).