Notes

INTRODUCTION: THE THREE REGIMES OF THE NOVEL

Epigraph: Mullan, How Novels Work, 9.

1. As You Like It III, 3; Fuentes, “In Praise of the Novel,” 614.

2. Blumenberg, “Concept of Reality,” 29.

3. Frege, “On Sinn and Bedeutung,” 57; I have replaced Sinn and Bedeutung, left un-translated for editorial reasons, with the customary “sense” and “reference.” (Bedeutung is occasionally rendered by “denotation” or “nominatum.”)

4. Pavel, Fictional Worlds, provides a capacious synthesis of theories of fictionality; he discusses early “segregationist” theories, which maintained a sharp line between fiction and factual discourse, on 11–17. Since the restrained historical definition of fiction offered in the present study has little common ground with the still-growing field of fictionality studies, for the most part I will not even attempt to canvas research on the philosophy of possible worlds, narratological approaches to the fiction-history distinction, and so on.

5. On Greek attitudes regarding the reality of gods and mythological heroes, see Veyne, Did the Greeks Believe; a major strand of thought held historical truth to be a kind of vulgate, consecrated by tradition.

6. Searle, Expression and Meaning, 72. (Searle includes maxims and the like as examples of nonfictional commitments.) Others have pushed in this direction; for a sample and critique, see Pavel, Fictional Worlds, 18–25.

7. De Man continues: “in the Iliad, when we first encounter Helen, it is as the emblem of the narrator weaving the actual war into the tapestry of a fictional object…. The self-reflecting mirror-effect by means of which a work of fiction asserts, by its very existence, its separation from empirical reality … characterizes the work of literature in its essence” (Blindness and Insight, 170). That Greek readers of Homer—at least by the time of Aristotle, to whom I will return in a moment—felt it to be a work of poetry as opposed to history seems to me indisputable; I doubt very much that such a status was predicated on a concept of the “empirical,” or an experience of the “essence” of literature.

8. “But are we quite sure we know what ‘literature’ means?” asks Roberto Calasso. “When we pronounce the word today, we are immediately aware that it is immeasurably distant from anything an eighteenth-century writer might have meant by it, while at the beginning of the nineteenth century it was already taking on connotations we quickly recognize” (Literature and the Gods, 170). Literature’s growth as a category has been analyzed by Caron, “Belles lettres.”

9. To my knowledge the lexical drift from “fiction” as synonymous with (usually devalued) poetic fancy to designating narrative literature as such has not been precisely traced. The titles of works like Staël’s Essai sur les fictions (1795) and Dunlop’s The History of Fiction (1814) at least suggest that by the turn of the nineteenth century “fiction” as an umbrella term for the novel exists in both French and English. For a brief look at the word’s English history, see Williams, Keywords, 134–35.

10. These scholars are Davis, Factual Fictions; Foley, Telling the Truth; and Gallagher, Nobody’s Story and “Rise of Fictionality.” Despite the sometimes stark disagreements that I develop below, my debt to these scholars, especially to Foley and Gallagher, is deep: they are the first to view “fiction” as a more problematic and interesting term than “the novel,” and this book would not exist without their work.

11. Walton, Mimesis as Make-Believe. For an additional philosophical-anthropological approach, see Schaeffer, Pourquoi la fiction? The argument from evolutionary biology is made in, e.g., Boyd, Origin of Stories.

12. As do Chevrolet, Idée de fable and Duprat, Vraisemblances. Richly detailed though they are, both studies take “fiction” as a historically unproblematic term.

13. See Blumenberg, Legitimacy. Blumenberg was countering the argument that the modern world is simply a secularized iteration of an earlier Christian world; it was, he held, legitimately new.

14. Poetics 9; in Russell and Winterbottom, eds., Ancient Literary Criticism, 102. (Further quotes in the following discussion are found here and on page 103.)

15. Wood, How Fiction Works, 238. Aristotelian poetry is thus understood as a hypothetical but logically coherent reality, as distinct from the purely fabulous and the historical: “Aristotle invoked, or invented, the concept of probability, in order to locate the reality of drama both beyond mere fiction and factual reality” (Pfeiffer, “Fiction,” 94–95).

16. Halliwell, Aesthetics of Mimesis, 195–97.

17. For “might,” see Aristotle, Poetics, trans. Hutton, 54; for “could,” see Aristotle, Poetics, trans. Halliwell, 40. For Aristotle’s poetry as a “hypothetical” or “imaginable” reality, see Halliwell, Aesthetics of Mimesis, 179, 188.

18. Halliwell, Aesthetics of Mimesis, 166. The interpretation of poetic generality as proto-fictional, though repeated, is not fully sustained in Halliwell’s account; my reasons for disagreement will become clear. For an alternate argument that the Greeks’ “poetry” had little to do with “fiction” despite occasional congruence between the two, see Gill, “Plato on Falsehood.”

19. Diderot, Discours sur la poésie dramatique, in Oeuvres esthétiques, 212. “Facts” here refer to people and actions, while Diderot uses “interest” in the sense common to aesthetic theory in eighteenth-century Britain and France—emotional appeal or impact. Poetic “additions” often include invented secondary characters who enrich the plot: as I will explain in Chapter 1, the point is not that Aristotelian poetics tolerates no (or only minimal) invention of characters, but that attested events and heroes are its point of departure.

20. For a detailed consideration of the Greek novels’ situation with respect to history, see Morgan, “Make-Believe,” and Heiserman, Novel Before the Novel. Bakhtin’s investment in uncovering “realistic” novelistic chronotopes leads him to emphasize the importance of works like the Satyricon at the expense of the “adventure-time” chronotopes of Greek novels (see Dialogic Imagination, 84–258). As I discuss briefly below and again at more length in Chapter 2, however, it is the latter works that had a more demonstrable influence on early European novelists, who nonetheless quickly abandoned the use of invented heroes so as to bring their creations in line with Aristotelian stipulations regarding the preeminence of historical subject matter.

21. In this vein, see Catherine Gallagher’s remarks on Shakespeare: “It is difficult to say how much Shakespeare or his audience were invested in the assumption that Hamlet had been a real person, but it does seem that the playwright had a much harder time than we do imagining that tragic heroes had no prior bodily existence” (“Novel,” 231n6). The gist of Gallagher’s remark seems to me right, even if, as will become clear, the issue is probably not best framed as one of imaginative ability.

22. The pioneering study is Tieje, “Peculiar Phase.” I will be mentioning a number of others.

23. Richardson, Selected Letters, 85.

24. Coleridge, Biographia literaria, 7. It should be noted that Coleridge, who in contradistinction to Richardson glosses this willingness as “poetic” (not historical) faith, is referring here to the reading of specifically supernatural subject matter, and not poetry about the everyday world (whose representation needs no “suspension of disbelief” at all). In other words, this famous phrase, in context, does not describe a phenomenology of reading (or of film- or theater-going) tout court, as is now routinely assumed. Unfortunately, many critics have rushed to use Coleridge’s formula as a statement of modern fictionality, the “realization” toward which earlier thinkers were tending; among many others, see Gallagher, “Rise of Fictionality,” 347–49, and McKeon, Origins of the English Novel, 128, 297.

25. Burke, Philosophical Enquiry, 47.

26. Burke is of course speaking of theatrical spectacle, and Richardson of the novel, but despite the qualitative difference between reading a book and seeing and hearing bodies on stage, the two types of consumption were continually conflated in the period.

27. Burke, Philosophical Enquiry, 47.

28. To be precise, two meanings of the term are overlaid in Richardson’s letter. When he writes of “that kind of Historical Faith which Fiction itself is generally read with,” fiction seems to mean approximately “literature”; when he adds “tho’ we know it to be Fiction,” fiction now means something untrue, unhistorical.

29. Quoted in Reiss, “Imagining the Worst,” 106–10. The Dorking episode is analyzed in Clarke, Voices Prophesying War, 27–57.

30. Balzac, Comédie humaine, 3: 50.

31. Balzac, Comédie humaine, 1: 11. As we will see in Chapter 3, manners (as opposed to, say, passions or surprising adventures) had been an object of earlier novelists, though the latter had spoken more of producing a “tableau” or “picture” than a “history.”

32. Balzac, Comédie humaine, 1: 11.

33. Todorov and Genette, eds., Littérature et réalité, 7. The only essay in the collection not to take such a position is the French translation of the first chapter of Ian Watt’s Rise of the Novel.

34. Barthes, “Effet de réel,” 89.

35. “In Barthes’s description, reality effects are designed to create the aura of real life through their sheer meaninglessness: the barometer [Barthes’s example of a superfluous object] doesn’t play a role in the narrative, and it doesn’t symbolize anything. It’s just there for background texture, to create the illusion of a world cluttered with objects that have no narrative or symbolic meaning” (Johnson, Everything Bad, 78). Similarly, according to James Wood’s gloss on Barthes’s concept, “[realist] fiction builds into itself a lot of surplus detail just as life is full of surplus detail”; the physical appearance of a character will be described in superfluous depth because “it is just ‘how he looked’” (How Fiction Works, 81).

36. Barthes, “Effet de réel,” 89.

37. For this critique, see Tallis, Not Saussure, and Compagnon, Démon de la théorie, 141–47.

38. On the pragmatic self-refutation of the argument for reference’s impossibility, see Tallis, Not Saussure, 59. Michel Riffaterre’s essay “L’Illusion référentielle” (also reprinted in the Littérature et réalité collection) avoids this absurdity by restricting nonreferentiality to literary language. But now the argument becomes self-confirming, as Compagnon points out: literary language is defined as nonreferential, and manifestly referential books or parts thereof are simply denied literary standing (Démon de la théorie, 130–31).

39. Prendergast, Order of Mimesis, 70–71.

40. This implied hallucination—Prendergast’s term—represents “a remarkable blind spot in the semiological approach to the question of reference” (Order of Mimesis, 71–72).

41. Compagnon, Démon de la théorie, 134. The relaxation of Barthes’s iconophobia in La Chambre claire—where the “indexical” reality effect of the photograph is positively described—is perhaps the result of contamination with the Lacanian idea of the imaginary. See Jay, Downcast Eyes, 437–56.

42. Barthes, “Effet de réel,” 89.

43. Barthes, “Écrire,” 25. Spectatorial naïveté did routinely pop up, however, in the many full-scale critiques of the realist novel appearing in the 1970s and early 1980s. The mendacity of the realist novel is driven home by Catherine Belsey with an updating of Stendhal’s anecdote: “The success with which the Sherlock Holmes stories achieve an illusion of reality is repeatedly demonstrated. According to The Times in December 1967, letters to Sherlock Homes were still commonly addressed to 221B Baker Street, many of them asking for the Detective’s help” (quoted in Tallis, In Defence of Realism, 153). In such attacks (of which Tallis mentions many), the self-reflexive novel emerges as the antidote, stepping in to warn readers about taking fiction for reality. (On infatuation with Sherlock Holmes as a more ironic phenomenon than Belsey thinks, see Saler, “‘Clap.’”)

44. Eliot, Adam Bede, 175.

45. “A True History” is incorporated into the title page of Behn’s purportedly eyewitness account, Oroonoko, or, The Royal Slave (1688).

46. For instance Balzac, who like many realists had an affection for the mirror analogy, explicitly invoked Leibniz’s idea of a “concentrating mirror,” or speculum concentrationis (Balzac, Écrits sur le roman, 59n1).

47. Balzac, Comédie humaine, 1: 14.

48. My remarks on fiction’s ability to make a specific type of proposition—subtle propositions that are worlds away from the naïveté lambasted by Barthes and others—are brief. For a more extensive treatment, see Foley, Telling the Truth, 42–63.

49. Dickens, Hard Times, 314.

50. Cited in Vargas Llosa, Temptation, 168.

51. Balzac, Écrits sur le roman, 83; Hugo, Oeuvres complètes, 14: 1254. The other quotations (made without reference to the formulations of Balzac and Hugo) are from, respectively, Furst, All Is True, 9, and Dubois, Romanciers du réel, 47. (Though both Furst and Dubois want to save realism from such accusations, they do so by balancing its mendacity against more positive characteristics; the accusations themselves are validated.)

52. For some foundational accounts in this vein, in order of initial appearance, see May, “Histoire”; Deloffre, “Problème”; Mylne, Eighteenth-Century French Novel; Coulet, Roman jusqu’à la Révolution; Showalter, Evolution of the French Novel; and Démoris, Roman à la première personne.

53. Stewart, Imitation and Illusion, 303. In “Rise of I,” Stewart has substantially modified his perspective, stressing the formal importance of first-person narration in a way that is largely compatible with my understanding, developed below, of the pseudofactual novel as a technology.

54. Cohn, Distinction of Fiction, 3.

55. Davis, Factual Fictions, 156.

56. Foley, Telling the Truth, 145, 144.

57. Foley, Telling the Truth, 118, 119.

58. Foley, Telling the Truth, 118.

59. Gallagher, Nobody’s Story, xvi.

60. Gallagher, Nobody’s Story, 145.

61. Gallagher, “Rise of Fictionality,” 344. I will return to Gallagher’s account of Fielding’s types in Chapter 3.

62. To be sure, the critic is careful not to suggest an instantaneous development: “there was no sudden novelistic revolution that purged English narrative of somebody [i.e., real “referents”] and replaced him or her with nobody [fictional characters]” (Nobody’s Story, 165). To square this prudence with the idea of a “massive reorientation,” perhaps Gallagher would argue that there is indeed a process, only a quick one, completed in the space of a couple of decades.

63. Foley, Telling the Truth, 117.

64. Montesquieu, Oeuvres complètes, 1: 131.

65. How Gallagher would answer such questions is unclear, partially because of the nature of her claims and partially because her work has evolved. I’ve noted that Nobody’s Story does not attempt to take the measure of pseudofactual conventions; moreover, while it suggests systematic upheaval, it consists largely of close literary readings of the themes of property and production, self-invention and naming, debt and credit, all of which are seen as (metaphorically) engaging the issue of fictionality. Gallagher’s chronology also proves quite elastic, since in a third study she has suggested that it is in the nineteenth century that a general valorization of doubt (emblematized by Marx’s analysis of commodity fetishism) allows novel readers to entertain propositions they do not literally believe in; at the same time, this culture of doubt (and thus fiction) is said to have its roots in Hamlet (Gallagher and Greenblatt, “Novel”). By this standard, like the bourgeoisie of the old joke, fiction too will always be rising.

66. Foley, Telling the Truth, 124; she calls the longevity of the mode “a wonder” (119). Davis remarks the “striking fact” that Richardson and other novelists “refused to concede that they were writing fictions,” proposing that this may be “because fiction was too limiting a concept for them” (Factual Fictions, 176, 192).

67. “The convention of the claim to historicity is becoming increasingly vestigial,” reads McKeon’s gloss on Richardson’s letter to Warburton (Origins of the English Novel, 414).

68. For simplicity I’ve focused my attention on the three scholars of the English novel who have done the most to foreground fictionality as a historical problem. I should mention a few more here, starting with Day, From Fiction to the Novel, who offers a capacious tour of English pseudofactuality that argues, albeit diffusely, that its strategies are an effort “to create realism” (189). (Day’s vocabulary does not map onto that of most, since “fiction” in his title refers to the outright romance fancifulness that preceded “the novel.”) Without foregrounding the term, Michael McKeon suggests that fiction emerges as a kind of middle way between naïve empiricism and extreme skepticism. Through a slow dialectical process, the novel teaches its readers that literature involves a special sort of credence: in language that recalls that of Davis and Foley, he writes that little by little “modern culture becomes sufficiently tolerant of artful fictions to pass beyond the bare recognition of their incredibility and to conceive the possibility of their validation in other terms” (Origins of the English Novel, 128). The most recent attempt to tackle the problem of the French pseudofactual is Herman, Kozul and Kremer, Roman véritable; the idiom here is different, since what rises is neither realism nor a kind of credence but the “legitimacy” of fiction. Finally, there are a number of studies of the eighteenth-century French and English novel’s relation with history that do not foreground fictionality (or the pseudofactual) per se: Gearhart, Open Boundary, Ray, Story and History, and Zimmerman, Boundaries of Fiction.

69. Though I address the issue in much more depth in Chapter 2, speaking of romance here demands that I clarify my use of the word novel. Like many literary historians, I use “romance” to designate the long narratives of the French seventeenth century—works from d’Urfé’s L’Astrée (1607–27) to Scudéry’s Clélie (1654–60)—that were modeled on Heliodorus’s Aethiopica. Unlike many, however, I treat romance as a form of the novel—a form that is morphologically distinct from the pseudofactual novel and the fictional novel. (The distinction involves not only length, but also, say, the number of characters and modes of narration; the distinction is not one of content or intrinsic plausibility.) I avoid, then, the historical and logical difficulties of determining just when romance becomes the novel at the same time I acknowledge the temptation—visible since the 1660s—to make the distinction. That is, there is indeed a (morphological) break between romance and the pseudofactual; but there is no break between romance and the novel. (If I could, I would reword the statement of John Mullan’s I use as my epigraph: “the” novel does not arrive in the eighteenth century, only a certain identifiable form of novel.)

70. See Morgan, “Make-Believe,” 177–78. (These are among the rare—and invariably derogatory—references to the writings that Bakhtin and many others view as ancestors of the modern novel.)

71. Outillage mental was a term coined by the pioneering historian of mentalités Lucien Febvre; see Chartier, “Intellectual History,” 18–22.

72. Gossman, “History and Literature,” 23. Gossman’s contention might be better rephrased by saying that history and poetry were both co-equal branches of rhetoric, though classifications such as these were hotly contested in the Renaissance.

73. Poetics 9, in Russell and Winterbottom, eds., Ancient Literary Criticism, 102–3.

74. For the Chinese situation, for example, see Lu, From Historicity to Fictionality.

75. See Patterson, Negotiating the Past, esp. 197–230.

76. Quoted in Nelson, Fact or Fiction, 46. Nelson bases his efficient synthesis on Weinberg, History, who concludes, “Most critics believe that the object [of imitation] must be a true one if credibility is to result” (1: 633). In McKeon’s account of the novel, such statements are taken for signs of the inroads made by the empirical spirit. Hence, observing that references to historical truth can already be found in Renaissance romance, he declares them to be attempts by a formerly fanciful genre “to adapt to epistemological revolution and to keep itself honest” (Origins of the English Novel, 56). But scorn for poets who speak of people and events that never were does not make the people who deliver the scorn predecessors of Locke; they are merely good followers of Aristotelian thought about poetry. McKeon briefly denies this, arguing that such thinkers “tended to read the Poetics through the spectacles of empirical epistemology” (53). The contention, which is never supported, follows from the preconception that only the invisible hand of empiricism can motivate the appeal to truth.

77. Fielding, Tom Jones, 371.

78. Nelson calls this development a “flight from fiction,” by which he means a retreat into ever more historical poses; see Fact or Fiction, 92–115.

79. See Heliodorus, Histoire aethiopique, 159–64.

80. Quoted in Esmein, ed., Poétiques du roman, 139–40. Their preface is often cited as a manifesto of neoclassical vraisemblance. Again, this was just a rationale, and some did not agree: Huet’s treatise De l’origine des romans (1670), perhaps the best-known early history of the genre, maintains that romance protagonists do not need to be historical in the manner of tragic characters: being of middling stature, the former could not possibly have come to the attention of readers; they are vraisemblable in the sense that they may be assumed to be historical (Origine, 445–46). This is not, however, an argument for open fictionality: as we will see in Chapter 1, the logic is still firmly Aristotelian in its postulate of historical possibility.

81. “L’esprit n’est point ému par ce qu’il ne croit pas” (Art poétique III, 50, in Boileau, Oeuvres complètes, 170). Boileau restates Horace’s oft-reprised dictum of incredulus odi: “Whatever you thus show me, I discredit and abhor” (Ars poetica, in Satires, 466–67). “Thus” referred specifically to overly horrible events (Medea’s murder of her children, the dinner Atreus makes of his brother Thyestes), but the line was widely taken as declaring the necessity of audience belief across the board, especially in the matter of supernatural phenomena.

82. On illusionism’s deep roots in Aristotelian thought of the Renaissance, see Hénin, Ut pictura theatrum. Hénin’s work demonstrates that the concern with illusion that many have viewed as typical of the Enlightenment thought about the arts and a properly modern “aesthetics” (e.g., Hobson, Object of Art; Marshall, Frame of Art) is in fact merely a development of much older commonplaces.

83. D’Argens, Lectures amusantes, 1: 52–53.

84. This is not to say that Aristotelian invention could not be made to serve topical ends: French neoclassical tragedy as a whole can and must be viewed in the context of pressing political questions of the day (conflicts between the aristocracy and the monarchy, problems of governance, and so on); and heroic romance allowed a writer such as Madeleine de Scudéry to offer extended instruction in aristocratic sociability at the same time she referred to heroes of the past. But I follow Foley’s claim that the pseudofactual effectively expanded the reach of the novel’s power to figure the contemporary world.

85. The first such narrators seem to appear in England in the late 1780s—in, say, Blower’s Maria, A Novel (1785) and Wollstonecraft’s Mary, A Fiction (1788). By the 1790s, a number of very different subgeneric contexts feature third-person texts having no “real world” narrator, among them Austen’s novel of manners, Scott’s historical novel, and Goethe’s Bildungsroman; the gothic novel alternates between pseudofactual found documents and a fictional third person. Further narratological considerations will be postponed until the Conclusion.

86. Acknowledging Gallagher, Poovey, Genres of the Credit Economy, has greatly expanded this line of inquiry.

87. Bender, “Enlightenment Fiction.”

88. Reiss, Discourse of Modernism.

89. Welsh, Strong Representations, 44–76.

90. Liu, “Power of Formalism,” 743. On the affinity in the humanities for arguments that “enchant” their object of study, see Schneider, Culture and Enchantment.

91. Kant, Critique of Judgment, 43. By contrast, Kantian pleasure is predicated on an object that really exists.

92. See Benjamin, Arcades Project, 150–70. Thanks to Andrew Clark for the Benjamin analogy.

93. To an extent, this is the kind of history of the novel attempted by Franco Moretti, especially in Graphs, Maps, Trees. I will have occasion to allude to Moretti’s findings in the course of this book, though usually because his hypotheses, couched in the Darwinian language of evolutionary biology, do not match my observations. Between what I’m calling a morphological history and Moretti’s work are two major differences. First, Moretti thinks of formal innovation as a kind of gene passed down from generation to generation, whereas the technology metaphor is, I believe, much less strained. Second, though Moretti crunches an impressive amount of brute data, the way he interprets the data is very much in line with the “sociology of forms” he has advocated in the past: observed evolutions must be the epiphenomenal manifestation of factors such as the rise of the bourgeoisie or capitalist market culture. (This is particularly evident in “Style,” his study of titles.)

CHAPTER 1. THE IMPOSSIBLE PRINCESS (LAFAYETTE)

Portions of this chapter appeared in “Lafayette’s Impossible Princess: On (Not) Making Literary History,” PMLA 125, 4 (2010), reprinted by permission of the copyright holder, the Modern Language Association of America.

1. Watt, Rise of the Novel, 30. The modernity of La Princesse de Clèves is proverbial. For some accounts, see Stone, “Exemplary Teaching”; Prince, Narrative as Theme (39–50); Lyons, Exemplum (217–36); and Desan, “Economy of Love.” For a review and critique of such arguments, see Campbell, “‘Modernité.’”

2. I refer to what were known as nouvelles historiques as historical “novellas” rather than “novels” solely to avoid associations with later historical novels by Scott and others. By “historical romance” I designate what was called the roman héroïque, or, as time went on, the vieux roman. This distinction was generic, and was observed at the time Lafayette was writing. My use of it here does not imply that I believe that the roman héroïque was not a “real” novel, or that the nouvelle historique was. Many things separate the two forms, but from the point of view of this study both are equally far from being fiction.

3. Charnes, Conversations, 84–85 (citations in the text will appear with the abbreviation C). Valincour’s most recent editor follows this argument; see Valincour, Lettres, 25–30 (henceforth abbreviated V). Assertions of Charnes’s superiority on this point have become commonplace; see, e.g., Esmein, ed., Poétiques du roman, 632–42.

4. Valincour’s criticisms are attributed in his text to different interlocutors; since we cannot know what Valincour’s own opinion was, I refer to all the (sometimes contradictory) critical statements as his.

5. Seventeenth-century French romances typically had some sort of historical setting. History both gave the works a claim to moral value (Bannister, Privileged Mortals, 91–99) and underwrote the verisimilitude responsible for readerly pleasure: “[when] lies are made openly, such crude falsity makes no impression on the soul, and gives no pleasure,” wrote Georges and Madeleine de Scudéry in the preface to their romance Ibrahim (1641–44; Esmein, ed., Poétiques du roman, 139–40). As I’ve noted in the Introduction, occasionally writers offered contrary opinions without developing a practice that could displace the Aristotelian model.

6. Valincour’s position here was representative; another contemporary formulation can be found in Sorel, De la connaissance des bons livres (1671) : “When the novel is entirely invented, it can be set up [réglé] according to the author’s fancy…. But if an author takes his subject from some ancient or modern history, he is not allowed to change the fundamental actions; one can certainly add to the truth, but not corrupt it, otherwise this would mean contradicting everything that historians had written about it” (Esmein, ed., Poétiques du roman, 319). Tragedy had been envisioned no differently, and it is instructive that Corneille’s pride over his limit case for poetic invention, Rodogune (1644–45), was based on the fact that he had managed to attribute to a cast of real characters a completely fabricated plot. (Invented characters were shunted off to the lower genres of tragicomedy and comedy.)

7. Genette, “Vraisemblance et motivation,” first drew attention to this aspect of Valincour’s critique, and feminist scholars in particular have concentrated on how female characters’ plausibility was seen to vary in proportion to submission to patriarchal convention; see especially Miller, “Emphasis Added,” and Beasley, Revising Memory. On cultural plausibility generally, see Culler, Structuralist Poetics.

8. Bussy’s objections center around the scene in which the Princess admits her temptations to her husband; Fontenelle’s reserves are much more muted, extending only to the conduct of Nemours and Monsieur de Clèves. Both are reproduced in Laugaa, Lectures, 18–19, 22–25.

9. On the importance of vulgarized history in the period, see Ranum, Artisans of Glory.

10. The most influential analyses of the historical novella have centered around this notion of secret history. For a general account, see Harth, Ideology and Culture, 129–79. Since secrets in the historical novella are usually affective in nature and concern women and their power to shape history, the genre has often been read as a subversion of official masculinist historiography; see especially Beasley, Revising Memory (190–243), and Grande, Stratégies de romancières (361–84). See also the remarks of Hipp, Mythes et réalités (132–94), who looks closely at novelists’ engagement with history in this period, though without mention of Lafayette’s exceptionality.

11. Quoted in Nelson, Fact or Fiction, 43.

12. As might be expected, over the course of the half-century separating d’Urfé’s pastoral romance L’Astrée (1607–27) and the advent of the historical novella in the 1660s, what we might want to call the “historicity” of characters varied. Typically—in the romances of the Scudérys, of La Calprenède—protagonists had historical credentials, while minor characters were merely plausible. In Lafayette’s own stab at the romance genre, Zayde (1670–71), the eponymous heroine was “invented,” certainly, but in the accepted manner: extant documentation said nothing about her. Early in the century, one finds more indifference to historical referents. D’Urfé, for instance, placed his shepherds in fifth-century Gaul but did not pretend that they were real figures; in this, he was in line with the reception of Heliodorus’s Aetheopica, which was not assumed to have historical sanction. However, in none of these cases do we find, as in La Princesse de Clèves, a heroine who is “visibly false” in the manner I will be describing.

13. For evidence that Lafayette’s readers were well versed in Renaissance history and would read with it in mind, see Letts, Legendary Lives. (Strangely, Letts does not comment on the fact that the novel’s heroine punctured a hole in this type of reading.)

14. Lafayette, Romans et nouvelles, 260 (henceforth abbreviated L). The supposed relation between Mademoiselle de Chartres and the historically real Vidame de Chartres (said to be her uncle), rather than authenticating her, would only have confused matters: “Vidame de Chartres” was a hereditary office, not a title of nobility or a family name (see Ritter, “La Philothée,” 88).

15. The matter of the number of de Clèves brothers is a bit intricate. It has long been thought that Lafayette’s mention of three brothers was due to her consultation of Anselme’s Histoire de la maison royale de France (1674), the only source to add a third brother, Henri de Clèves, to the two attested in other histories, François II de Clèves and Jacques de Clèves. (For this hypothesis, see Chamard and Rudler, “Sources historiques,” 125.) Valincour, clearly aware of Anselme, confuses Jacques and Henri, who is the one who never married (cf. Anselme, Histoire, 1:288). Seemingly unaware of Anselme, Charnes maintains that Lafayette simply invents a third brother (of the safe “possible” variety). Weil has proposed that because Lafayette may have finished her novel before the appearance of Anselme’s genealogy, Charnes is perhaps closer to the truth (C 108). It is difficult to imagine Lafayette passing up the opportunity to consult Anselme’s new volume, whatever the date of the work’s composition, but the objection changes little: she pointedly makes her protagonist the second-born son, even though sources all give this role to the married Jacques.

16. Brantôme preserves moreover the sexual associations of the colors, which symbolize “pleasure and steadfastness, or steadfast pleasure” (jouissance et fermeté, ou ferme en jouissance; Oeuvres complètes, 3: 271).

17. Indeed, all contradictions with respect to history were not equal. Dates, for example, which Lafayette used with considerable latitude, did not attract Valincour’s attention, as was to be expected in a literary culture in which dramatists regularly compressed events (see, e.g., Racine’s prefatory comments on Mithridate [1673]).

18. Compare, for instance, to Huet’s separation of history and romance: “[Histories] are true in the main, and false only in a few parts; romances, on the contrary, are true only in parts, and false in the main” (Origine, 445).

19. Harth, Ideology and Culture (194–95), has suggested that the historical novella routinely plays off virtuous fictional heroines against historical male figures, but this is misleading: only in rare cases do invented characters play significant roles, and even then they remain within the bounds of the possible (i.e., they do not contradict the historical record). The Appendix to the present chapter details the practice of Lafayette’s competitors.

20. I have not attempted to coordinate the above account with three other readings that have stressed the novel notion of fictionality at work in Lafayette because the terms of discussion are so different. But I would like to mention these readings here. In a deconstructive analysis by Mireille Calle-Gruber, the text is said to expose the workings of the fictive that it simultaneously puts into place (Effet-fiction, 169–200); Éric Van der Schueren holds the Princess’s fictivity to be a consequence of the Jansenist theory of the sign (“Portrait”); closer to my own preoccupations, William Ray has considered the fictional heroine as a “private” challenge to the collective discourse of history (Story and History, 24–49). None of these readings focus attention on the period’s conventions of poetic invention and novelistic practice, or their implications with respect to a history of fictional modes.

21. The denial is explicitly read as an ancestor of modern legal disclaimers by Noille-Clauzade, “Considérations logiques,” 181.

22. Indeed, assertions of pure invention were an unmistakable invitation to look for real-world referents. Keys thus provided an anchor for characters with no obvious historical or mythological precedent. Although such a practice did part ways with Aristotelian verisimilitude (functioning rather more along the lines of satire), it was no more “fictional” than Valincour’s idea of poetic invention. (On the many uses of keys in the period, see Bombart and Escola, eds., Lectures à clés, and Chapters 2 and 3 below.) In the particular case of La Princesse de Montpensier, the notice’s use of the word fabuleux (“imaginary”) would have been especially suspicious: the term, typically an antonym of “history” (see Sermain, Métafictions, 21–84), was out of place in a book whose use of history was on display; the notice thus generated a contradiction that could be resolved only by concluding that irony was intended. Nonetheless, the work is not keyed in any normal sense of the term: the roman à clef flagrantly depicts scandal under assumed names, while Montpensier uses real names to hint at impropriety.

23. Montpensier would on the face of things be a potential mentor of Lafayette, having included pieces by the renowned young wit in the anthology of portraits she had privately printed in 1659, thus giving Lafayette her first publication. On the other hand, allegiances were complicated: Lafayette was fast friends in these years with the scholar Gilles Ménage (who saw to it that La Princesse de Montpensier would be issued by his normal publisher, Thomas Jolly), and Ménage may have encouraged Lafayette to take a swipe at the official protectress of his own enemy, the Abbé Cotin (Duchêne, Madame de Lafayette, 201–3). It is also possible, as Micheline Cuénin has noted, that an affair between Guise and Renée d’Anjou, whose memory seems to have been expunged from the family records, was a matter of oral speculation in the entourage of the Grande Mademoiselle (see Cuénin’s presentation of the text in Picard and Lafond, eds., Nouvelles du XVIIe siècle, 1383). Marc Escola, the most recent editor of La Princesse de Montpensier, has presented the fullest case for Lafayette’s possible indiscretions; see Nouvelles galantes, 37–49.

24. The inverse relationship obtaining between the two characters may be signaled obliquely in a number of the novel’s details. I will limit myself to one example. La Princesse de Clèves’ mention of the husband of Lafayette’s former heroine, the prince de Montpensier, is hardly innocent: in La Princesse de Clèves, Montpensier is none other than the first candidate to marry Mademoiselle de Chartres. (The plans go quickly awry due to the duchesse de Valentinois’s opposition.) Mademoiselle de Chartres, then, is the young woman who barely avoids the fate of Renée d’Anjou. This type of detail is admittedly of little interest to the modern reader, but might not have gone unnoticed by a public fond of genealogical intricacies.

25. Joan DeJean, “Lafayette’s Ellipses,” was the first to see in the novel’s gossip something other than a mirror of court life: when read alongside Lafayette’s documented reluctance to identify herself as the work’s author, La Princesse de Clèves starts to look like a comment on the period’s abundant literary commerce in stories of feminine passion. For the proposal that the novel be read as a comment on keyed reading specifically, see Gevrey, “Lectures à clés.”

26. On such naming, see Horowitz, “Primary Sources,” 166; somewhat in the manner of DeJean, Horowitz reads the withdrawal of the Princess at the novel’s end as “resonant testimony to the power of the not named” (175).

27. In two of the novel’s interpolated narrations Lafayette does recount gossip—the loves of Diane de Poitiers and Henry II, and of Anne Boleyn and Henry VIII. Such stories had in a sense passed over into the public domain, which is why Fontenelle wonders how it would be even possible for the fifteen-year old protagonist not to have heard them already. “Even I knew them,” he objects (quoted in Laugaa, Lectures, 25).

28. Whence Lafayette’s advantage over even one of her most talented contemporaries, Villedieu, whose moralistic denunciation of the treachery of court and heart in Les Désordres de l’amour was more part of the problem, as the saying goes, than its solution. For Villedieu publishes in her third novella an archival “scoop”—the real parting love-letter actually addressed by the protagonist Givry to his unreceptive mistress shortly before his death. See Villedieu, Désordres, 206.

29. In his account of gossip in Lafayette, Ross Chambers also proposes that the text “transcend[s]” the mundane phenomenon it describes. His argument proceeds differently, however, since he holds gossip and novelistic storytelling to be variations on the same thing: literature maintains “a complicity with the force of desire that is nevertheless the scandal of an orderly bourgeois society,” and La Princesse de Clèves in particular takes as its “enunciatory model” the characters’ own interest in each other’s private affairs (“Gossip and the Novel,” 213, 218, 224). The novel’s superiority to gossip, for Chambers, lies in its ability to provide detached psychological insight in the place of mere facts. However, the most basic reason the novel is superior to the gossip exchanged by its characters is that Lafayette’s subject, the Princess, has no historical referent.

30. See Hampton, Writing from History.

31. On the “contagion” model of aesthetic effect, see Chapter 4 below.

32. Occasional assertions are made to the contrary (e.g., Malandain, “Écriture de l’histoire,” 20), but there is no evidence to back up such assumptions. Alain Niderst, the foremost proponent of biographical keys to French texts of this period, has proposed a number of possible “models,” concluding that “the Princess is not imaginary”; they are entirely of his own devising, however, and not attested in the reception history (L xxxvi–xxxvii). By contrast, contemporary allusions to La Princesse de Montpensier all treated it as either keyed or an example of “brazen calumny” (Villiers, Entretiens, 152–53). Ironically, the sole allusion to the reality of the Princesse de Clèves was made by none other than Lafayette herself, in her oft-cited remark to the effect that she had heard that it was not a novel but in fact memoirs whose title had been changed; for the text of her letter, see Laugaa, Lectures, 16n1.

33. Historical novellas that were taken to allude to present-day affairs set their narratives in the present, most famously in Du Plaisir’s La Duchesse d’Estramène (1682); and most are not shy about advertising their topicality by stating at the outset that the story is a true one.

34. See Jeanneret, “Commentary.”

35. Most readings of the Mercure galant campaign stress its novelty; Joan DeJean for example has argued that it was instrumental in the formation of a new, expanded, non-coterie public for literature (Ancients, 51–66), and I have made similar remarks myself (“Storyteller,” 157–61). Yet notwithstanding the new forum the Mercure galant provides, Donneau de Visé’s prescribed reading harks back to a use of character that the novel itself no longer practices.

36. Valincour is incapable of seeing, writes Charnes, “that what he condemns [in one place] derives naturally from what he admired before” (C 172–73).

37. “Insidiously, Valincour throws doubt on the originality of the avowal,” writes Laugaa (Lectures, 91), while Cuénin gives credit to Charnes for rescuing Lafayette from unjust charges of plagiarism when he claims that La Princesse de Clèves had been completed before Les Désordres (Villedieu, Désordres, 231–32).

38. Allusions to the novel’s psychological exploits are too numerous to cite; but on Lafayette’s cultivation of a space of verbal discretion and “recessive action,” see François, Open Secrets, 66–128.

39. “Hesitatingly in the works of Defoe, then more boldly in subsequent novels of the eighteenth century, the pseudofactual imposture signaled the invocation of a mimetic [i.e., fictional] contract,” writes Foley (Telling the Truth, 118). For a more thorough presentation of such histories of fiction, see my Introduction.

40. Gallagher, “Rise of Fictionality,” 344.

41. Moretti, Graphs, Maps, Trees, 76.

42. In fact, Benjamin’s text reads: “It has rightly been said that all great works of literature found a genre or dissolve one—that they are, in other words, special cases”; his rephrasing of the sententious statement may well be a subversion of it, since it makes of the masterwork something atypical, as befitting an essay on Proust, in whose work “everything transcends the norm” (Illuminations, 201).

43. For a critique of this assimilation of natural and cultural evolution, see Prendergast, “Evolution and Literary History”; Moretti responds in “End of the Beginning.”

44. “Rousseau stimulated … in his anonymous readers a process of identification that was without real precedent,” writes Vincent-Buffault (Histoire des larmes, 14), taking no note of Valincour’s precedence. I will take up the problem of Rousseauian identification in Chapter 4.

45. See, e.g., the remarks of DeJean, Ancients, 181n74, and Fournier, Généalogie du roman, 290. Camille Esmein, “Pensée du roman,” has suggested that Valincour’s appreciation is part of a broader late seventeenth-century reorientation of novelistic reading practices in the direction of identification. Her richly documented study does not, however, do enough to distinguish the rare identificatory language of self-recognition (as seen in Valincour) from a much more widespread vocabulary of shared feeling and illusion.

46. See Laugaa, “Réception,” 121–32.

47. I explore the contemporary devaluation of the long romances at more length in Chapter 2.

48. It is possible that Lafayette had in fact composed (or at least planned) her work before the historical novella had been named and become popular: in 1671, a royal privilège is taken out for a book entitled Le Prince de Clèves. If this is the case, she could obviously not break rules of a genre that did not exist yet. But since that genre’s rules regarding invention were essentially the same as the rules of historical romance, tragedy, or any serious genre, the possibility would not change my point substantially.

49. Davis, Factual Fictions.

50. See Cohn, Transparent Minds.

51. I have culled my inventory of historical novellas from the all-but-comprehensive list of novelistic works provided by Lever, Fiction narrative. I rely on Lever’s research for attribution of these frequently anonymous tales.

52. Quirks include Préchac’s Le Secret, nouvelle historique (1683), which is really a collection of exempla from past and present on the necessity for secrecy, or Mailly’s late Les Disgrâces des amants, nouvelle historique (1697), in which the term, now a quarter-century old, is applied to a present-day amorous adventure with no historical ramifications whatsoever.

53. If Zayde has never been classified as a historical novella, this is because of its prominent use of interpolated tales, which clearly refers it to Scuderian romance; if one considers only its “historicity,” it is indistinguishable from many historical novellas, and indeed more detailed than a good number of others.

54. Two curiosities, though, actually borrow Lafayette’s invented characters: Bernard’s Le Comte d’Amboise (1688) reprises the dissimulating Madame de Tournon and the long-suffering Sancerre; Mademoiselle de Jarnac (1684), attributed to Boisguilbert, gives a supporting role to none other than the princesse de Clèves.

55. Harth, Ideology and Culture (194–95), adduces two texts as evidence for her contention that historical novellas routinely feature virtuous fictional heroines played off against underhanded historical male characters—Durand’s Histoire des amours de Grégoire VII (first published in 1687) and Courtilz’s Le Grand Alcandre frustré (1696). But the former has an eleventh-century setting that makes the invention both palatable and plausible, while the latter is not a historical novella at all, since it is, with the exception of the heroine, keyed to the contemporary loves of Louis XIV. The closest analogue to La Princesse de Clèves may in fact be a book whose male hero provides the title: in d’Aulnoy’s Histoire d’Hypolite, comte de Duglas (1690), the protagonist is grafted onto a known family, and available sources would have easily enabled readers to detect the grafting. (L’Histoire d’Hypolite’s use of history is analyzed extensively in Duggan, Salonnières, 165–200.) Still, exceptional as it too may be, L’Histoire d’Hypolite upsets neither my specific contention in this chapter (that Lafayette is the only writer to engineer a collision between her novel and history) nor my larger point (inventing protagonists is not a practice that spreads in the wake of La Princesse de Clèves).

CHAPTER 2. QUIXOTE CIRCA 1670 (SUBLIGNY)

An earlier version of portions of this chapter appeared in “Relearning to Read: Truth and Reference in Subligny’s La Fausse Clélie," in The Art of Instruction: Essays on Pedagogy and Literature in Seventeenth-Century France, ed. Anne Birberick (Amsterdam: Rodopi, 2008), reprinted by permission of the publisher.

1. The original 1670 Paris edition, attested in at least one early source, is lost; three Dutch editions survive (1671, 1672, 1680), though the 1680 edition, and perhaps even that of 1672, may be a repackaging of unsold volumes of the earlier printing. An English translation (1678) and the spate of early eighteenth-century editions (1710, 1712, 1716, 1718) do, however, confirm that the work did not sink into oblivion.

2. For French comic novels reprising Cervantes’s device, see Serroy, Roman et réalité; the stage (and much else) is covered by Bardon, Don Quichotte en France, esp. 1: 167–209.

3. Chapelain’s letter is quoted in Esmein-Sarrazin, Essor du roman, 11. La Calprenède died in the middle of his Faramond, which was finished by Ortigues de Vaumorière in 1670. On Scudéry’s transition, see Grande, “Du long au court.”

4. Du Plaisir, Sentiments sur les lettres, in Esmein, ed., Poétiques du roman, 761. Sixty years later, d’Argens was one who restated the position: “People tired of those long romances [romans]” (D’Argens, Lectures amusantes, 1: 24). (This is only one position, however: Lenglet Du Fresnoy, in his 1734 De l’usage des romans, registers no such lassitude, nor any radical break.)

5. Behn, Oronooko, 9; Manley is quoted in McKeon, Origins of the English Novel, 60.

6. The French routinely referred to both old and new novels under the umbrella term of roman; though it may have seemed, in the late seventeenth century, that the term nouvelle could replace it, this did not occur (Sgard, “Mot ‘roman’”). (For some guides to the confusing French terminology of the latter part of the seventeenth century, see Showalter, Evolution of the French Novel, 11–37, and Esmein-Sarrazin, Essor du roman, 36–43.) In England, the cognate terms “romance” and “novel” did eventually stabilize, thus reifying an opposition that is thematic (one is silly, one is realistic), formal (one is long, the other short), and diachronic (one replaces the other). The stabilization was slow, however: “seventeenth- and early eighteenth-century writers often use the terms ‘romance,’ ‘history,’ and ‘novel’ with an evident interchangeability that must bewilder and frustrate all modern expectations,” writes McKeon after reviewing booksellers catalogs (Origins of the English Novel, 25). And of course the fact that stabilization finally occurred in England but not France has not kept researchers from seeking the origins of a qualitatively new (modern) French novel. But is the romance-novel opposition really the best way to understand the period’s admittedly common urge to draw a line between old and new forms? Such is the question that will preoccupy me here.

7. See, e.g., Godenne, “Association,” esp. 73–74.

8. McKeon, Origins of the English Novel, has devised an elaborate “dialectical” model of generic evolution to explain these apparently regressive movements. I hope to show, however, that what seems to be retrograde motion is an effect of viewing literary history from the unmoving earth that is the “modern novel.”

9. Cervantes, Don Quixote, 28 and elsewhere. I quote from Smollett’s translation, which is exact: Cervantes always writes “books” (libros), and Spanish possessed no word corresponding to “romance” (roman and romanzo in French and Italian).

10. See, e.g., Quint, Cervantes’s Novel.

11. For one formulation of this position, see McKeon, Origins of the English Novel: “the epistemological reversals of Don Quixote, and Part I’s hypostatization under the self-conscious scrutiny of Part II, constitute an elaborate mechanism for inducing that species of belief-without-really-believing which would become, once the mechanism itself proved unnecessary, the realm of the aesthetic” (282).

12. On the French comic novel as an instance of symbolic class warfare, see Harth, Ideology and Culture, 34–67, and DiPiero, Dangerous Truths, 23–61, 163–93. There is much to be said for this version of events, though the professional situation of writers such as Sorel, Scarron, and Furetière (and no doubt Subligny) can be quite complicated. Sorel especially has attracted much interest; see, e.g., Stenzel, “Discours romanesque,” and Giavarini, “‘Histoire véritable.’”

13. The puzzle, which is not of my own invention, is most often recognized by scholars trying to square Don Quixote and Persiles. “After symbolically burning the chivalric romances he struggled to renounce, Cervantes began again and died writing a last chivalric work” (Robert, Old and the New, 3); “Why would Cervantes—so vocal and assiduous about toppling romance in Don Quixote—become engaged, toward the end of his life, in kidnapping it back again?” (Wilson, Allegories of Love, 4). Wilson recognizes what Robert, accepting the contradiction as such, does not: that Persiles is in no way a “last chivalric work.”

14. Doody, True Story; the premise (not argument) of Doody’s archetypal study is that the novel and romance are one and the same thing.

15. For remarks on the earliest uses of the Old French romanz, especially in its differentiation from the chanson de geste, see Marichal, “Naissance du roman.”

16. Amyot’s “Proesme” can be found in Heliodorus, Histoire aethiopique, 159–64; for his remarks on narrative structure, see 160–61. A wider context for Amyot’s praise of Heliodorus’s ordo artificialis specifically is offered by Cave, Pré-Histoires, 129–41. On the more general congruence between the 1547 preface and humanist poetics, as well as an account of the Aethiopica’s influence on French romance, see Sandy, “Classical Forerunners”; similarly, for Amyot as background for Cervantes’s Persiles, see Forcione, Cervantes’ Christian Romance, 13–29. Also valuable is Plazenet, Ébahissement.

17. Subligny, Fausse Clélie, 2 (henceforth cited parenthetically).

18. For consistency and brevity, and except in cases where confusion might result, I will refer to all characters by their last name.

19. “Esprit fort is a kind of insult directed at libertines and atheists incrédules] who think that they are above common beliefs and opinions” (Furetière, Dictionnaire universel [1690], art. “esprit”).

20. Montal’s association of superstitious discourse with socially inferior sources is typical of an inversion in the Western history of the marvelous, by which phenomena that had been the subject of elite interest become stigmatized as popular error; see Daston and Park, Wonders.

21. Thomas Pavel’s description of L’Astrée holds for romance more generally: “With the exception of a few cases of intentional deception, the stories of L’Astrée … are told by truthful and scrupulous narrators with perfect knowledge of events. These stories, moreover, are almost never challenged. They do not present the point of view of one character on the facts—a point of view that could be contradicted or at least modified by different testimony” (Art de l’éloignement, 251). This romance reliability had previously been pointed out by Horowitz, Honoré d’Urfé, 128.

22. That is, before approximately 1630, romances included a high proportion of first-person narratives; for reasons of aristocratic decorum, later authors would modify this procedure. For more information on romance narrators, with special attention to early modern transformations of the topoi of the Greek novels of Heliodorus and Achilles Tatius, see Plazenet, Ébahissement, 597–624.

23. On the truth topos in Marguerite de Navarre, see Mathieu-Castellani, Conversation conteuse, 7–22.

24. Shapin, Social History of Truth.

25. Poetics 9; quoted from Russell and Winterbottom, eds., Ancient Literary Criticism, 102.

26. Serroy, Roman et réalité, 672. Serroy seems to be following Coulet, who states: “Subligny engages in one single plot a group of characters whom we come to know by their behavior over the course of the action, by the stories they tell about themselves or that others tell about them, and by the commentary that these stories provoke” (Roman jusqu’à la Révolution, 254). As I have mentioned, in late French romances decorum prevented aristocratic characters from providing accounts of their own heroic actions; secondary characters—attendants and so on—instead provided the stories of their masters. The elimination of the first person was especially thorough by Scudéry’s time; of the 39 embedded narratives of Artamène, only 3 are in the first person (Godenne, Romans de Scudéry, 97n1).

27. Koyré, Closed World.

28. See Morlet-Chantalat, Clélie. Clélie’s use of commentary stands out quantitatively but not qualitatively from what was already found in L’Astrée or the Heptaméron.

29. As Morlet-Chatalat has remarked, Scudéry never pursues the possibility of interweaving her various narrative threads: between embedded tales and frame tale there is at best “a vague relation of anteriority that … never provides any surprise in the later events of the novel” (Clélie, 236). A better model for Subligny’s interweave might be the extremely intricate structure of L’Astrée or the enigmatic story of Destin and L’Étoile in Scarron’s Roman comique (1651–1657), although it, like Clélie, is exempt from worries about narrative truth. Subligny’s interest in the interweaving of narrative threads is shared by some of his contemporaries, notably, Lafayette, in Zayde (1670–1671), and Villedieu, in Les Exilés de la cour d’Auguste (1672). Challe’s Les Illustres françaises (1713) is sometimes seen as the last and best example of this sort of romance structure. None of these works, however, contain as much excess information as La Fausse Clélie, whose obsession with plotting connections overwhelms its own plot.

30. Coulet, Roman jusqu’à la Révolution, 254.

31. See, e.g., Serroy, Roman et réalité, 673–74.

32. Serroy, Roman et réalité, 672.

33. Again Serroy is surprised: “Juliette has only one fault, that of having lived a life so romance-like that it was indeed worthy of having figured in a precious novel [i.e., by Scudéry]” (Roman et réalité, 672).

34. A presentation and reproduction of Subligny’s preface can be found in Esmein, ed., Poétiques du roman, 352–55. Subligny might be responding here to a wish expressed by one of the devisantes in Segrais’s Nouvelles françaises (1656–1657), where the question of names is foregrounded: “I have a hard time believing that so many people of sense, the very ones who have imagined on our behalf such noble Scythians and generous Parthians, have not taken equal pleasure in imagining French knights or princes who would be just as accomplished and whose adventures would be no less amusing” (1: 19).

35. See, for example, Showalter, Evolution of the French Novel, who discusses Subligny’s naming practice specifically on 164–65. For some general remarks, see also Watt, Rise of the Novel, 18–21.

36. The Mercure was not the first successful periodical in France, but unlike the Gazette of Renaudot, which starting in 1631 featured matter-of-fact reporting of international and national events, it offered society news in a format that encouraged reader submissions and involvement. As such, it was very much the ancestor of Addison and Steele’s famous Spectator.

37. In contrast with the Heptaméron, in which “true” tales feature at best characters whose names are onomastic plays on character traits; see Winn, “Clin d’oeil.” This is not to say that scholars have not made efforts to identify—with apparent success—many of the real-world referents of Marguerite de Navarre’s work; see Cazauran, Heptaméron, 30–33.

38. Niderst, Essai d’histoire littéraire, 54. That such a decoding was well within contemporary expectations is clear from the fact that anagrammatical reading was practiced both in erudite circles (Hallyn, “Anagramme”) and in salon literature (Denis, Parnasse galant, 225–27).

39. The example consulted is in the collection of the Bancroft Library, University of California, Berkeley (PT1100.F7 no. 1061 \t\). Niderst to all appearances was unaware of this key when he made his identifications of Subligny’s protagonists.

40. Notwithstanding Niderst’s many correct guesses, most identifications given by the key do not work through onomastic resemblance. Montal is given not, as Niderst hypothesizes, as François de Montsaulvin, “second born son of the famous Comte de Montal, who fought alongside Condé after the Fronde” (Essai d’histoire littéraire, 54), but as the Chevalier de Cavois.

41. The heroine of Lafayette’s La Princesse de Clèves, made “impossible” in the manner I’ve detailed in Chapter 1, should be considered a reaction to the type of narrative exchange that La Fausse Clélie strives to teach its readers.

42. The most notable sign of a change in critical winds has been the appearance of a volume of Littératures classiques devoted to the subject (Bombart and Escola, eds., Lectures à clés). For pioneering work on the subject, see Beugnot, “Oedipe et le sphinx,” who argues that the significance of keys far outstrips the “rudimentary sociology” (229) often practiced by their nineteenth-century champions; and Stewart, “Roman à clefs,” who provides a brief overview of the post-romance use of keys in France.

43. Cousin, Société française.

44. For instance, the key covered too few of the book’s many characters to authorize us to view the entire opus as keyed. (For a critique of Cousin’s position, see Godenne, Romans de Scudéry, 83–96.) Since Cousin, other researchers have attempted to devise keys from scratch, based on various biographical clues; see especially Niderst, “Sur les clefs.”

45. Cited in Godenne, Romans de Scudéry, 88.

46. In her correspondance, Lafayette refers on two occasions, in 1657 and 1658, to Clélie’s keys (Lafayette, Oeuvres complètes, 544, 566); the Dutch scientist Christian Huygens refers to keys in his journals (see Mesnard, “Pour une clef,” 371; Mesnard’s article offers a spirited defense of the relevance of keys to Scudéry’s novels). The abundantly testified propensity of contemporaries to view Molière’s protagonists as having specific real-world models or the printed keys to La Bruyère’s later Les Caractères are additional evidence that the practice was widespread. (In the following chapter I will return to the problem of keyed readings in novels inspired by works of social observation such as La Bruyère’s.)

47. Scudéry, Clélie, 1: 496–501. One might also note that Scudéry imports the burgeoning genre of the literary portrait, which patently depends on real-world referents, directly into a novel whose eponymous protagonist’s very name (Clé-lie) subliminally spurs us, perhaps, to look for keys.

48. “It would therefore seem that while the portrait increasingly exhibits its link to the present, intercalated stories, with their narrative content, remain beholden to a tradition of moral reflection” (Morlet-Chantalat, Clélie, 214). Additional information on the portrait craze can be found in Harth, Ideology and Culture, 68–128, and Plantié, Mode du portrait.

49. On the period’s practice of pseudonyms, see Denis, Parnasse galant, 189–235.

50. Rohan was a brilliant member of court society who has the distinction of being the only member of the upper nobility executed during Louis XIV’s long reign (in 1674, after a failed plot against the crown).

51. Cervantes “could not have written Don Quixote at all without a keen sense of the difference, and the relationship, between what we now think of as ‘romance’ and the ‘novel’” (Riley, Don Quixote, 11). The proleptic slip is especially telling when it comes from the pen of a scholar who has done so much, in Cervantes’s Theory of the Novel, to show the congruity between Cervantes’s practice and earlier humanist reflection on romance. And knowledge of the anachronism does not seem to keep critics from indulging in it: “During the four centuries since the publication of Don Quixote there has been no doubt that to write a major work of fiction was to write a novel. And while Cervantes could not possibly have known that in writing Don Quixote he was about to initiate the most important literary genre of the modern age, it can retrospectively be said that, among novels, Don Quixote was the first” (Cascardi, “Don Quixote,” 58).

52. Amyot, “Proesme,” in Heliodorus, Histoire aethiopique, 159. Marc Fumaroli, “Jacques Amyot” (30), has pointed to this sentence as the germ of Don Quixote.

53. For a detailed presentation of the French Amadis and des Essart’s involvement, see Simonin, “Disgrâce d’Amadis.’”

54. Yates, Astraea, 108.

55. “For us, romance must always be romance revival,” writes Ian Duncan, associating “us” with writers from Scott forward (Modern Romance, 7). In fact, the statement holds even in the Renaissance.

56. Amyot, “Proesme,” in Heliodorus, Histoire aethiopique, 159.

57. For these contentions, see respectively Plazenet, “Jacques Amyot,” 238; Cappello, “Discours”; Fumaroli, “Jacques Amyot,” 34–39; and O’Connor, Amadis, 22, 242–45.

58. The influence of the Orlando on the Quixote has been well explored; see Brownlee, “Cervantes as Reader of Ariosto,” and Hart, Cervantes and Ariosto. It might also be recalled that subject matter actually migrated from the Orlando to later volumes of Amadis (Cioranescu, L’Arioste en France, 1: 360–62): the frontier between “bad” chivalric romance and “good” Renaissance works by Ariosto and Cervantes is largely artificial; the polemic hostility of a figure like Amyot masks a literary-historical reality that was considerably more fluid.

59. Montreux’s work was routinely taken for an authentic specimen, though Huet, deploying some serious historical and philological work in his De l’origine des romans (1670), detected the ruse. Fabricating Greek novels from scratch became something of a pastime for men of letters; see Sandy, “Classical Forerunners.”

60. Forcione, Cervantes’ Christian Romance, 3. Forcione continues: “The years following its posthumous publication in 1617 witnessed ten editions, translations into French, Italian, and English, and imitations in prose fiction and drama. In the eighteenth century new editions, imitations, and translations appeared, and in the early years of the nineteenth century a scholar of the stature of Sismondi could still affirm that many readers considered it to be Cervantes’ masterpiece.” Most specialists of the Persiles resolve its “contradiction” with Don Quixote in roughly the manner I have done here; in addition to Forcione, see Wilson, Allegories of Love. The problem is that these lessons have not been widely absorbed by historians of the novel. Finally, it should be noted that the idea of a contradiction between Cervantes’s works has an unexamined history, one that may well date from the tail end of Persiles’s popularity: “What shall we say of a man who had produced Don Quixote and could afterwards write a book of the same kind he satirized?” asked a participant in Reeve’s dialogue The Progress of Romance (59).

61. As I’ve mentioned in the Introduction, Amyot’s praise for Heliodorus’s plotting was counterbalanced by the assumption that since it did not feature real epic heroes, it was good only for diversion. In the French seventeenth century, the prestige of romance was assured by attaching it to history; this is why La Calprenède actually resists calling his Faramond a roman, saying that he is really offering “histories embellished with some invention” (“Au lecteur” [1661], reproduced in Esmein, ed., Poétiques du roman, 224).

62. Richelieu’s remark is reported in the introduction to John Davies’s 1657 translation of L’Astrée, and quoted by Doody, True Story, 269.

63. Huet, Origine, 533. On L’Astrée’s complicated civilizing function, see the foundational work of Elias, Court Society, 279–305, and the synthesis of Wine, Forgotten Virgo, 109–42. One might note that Amadis had already been vaunted for its civilizing effects; Sidney remarks on men who, reading the romance, “have found their hearts moved to the exercise of courtesy, liberality, and especially courage” (Apology, 114).

64. On the havoc absolutism played with the model of sociability espoused by Scudéry’s romances, see Goldsmith, Exclusive Conversations, 59–72; on the creation of something like a “mass market” for books, see DeJean, Reinvention; for a theorization of “leisure” literature, see Cherbuliez, Place of Exile; on the tendency of (sub)genres to change with generational shifts, see Moretti, Graphs, Maps, Trees, 3–30. This last issue is perhaps banal only in appearance: literary historians prefer to find epochal motivations for morphological change, but literary fashions may in some cases come and go with no more deep logic than, say, baby names or hula-hoops.

65. I myself have adopted this perspective; see Lafayette, Zayde, 15–16.

66. Huet, Origine, 442.

67. Subligny’s own denial of reference is just such a patent feint: “I would humbly request that those whose names may have some resemblance to the ones I invented refrain from imagining that I have done so on purpose. They will easily see by the lack of conformity between their adventures and the ones in my stories that it is more chance than intent. And at any rate, any gallant ladies and gentlemen there may be are always presented to their advantage” (n.p.).

68. Adam, ed., Romanciers du XVIIe siècle, 901.

69. I am ventriloquizing the frustration of Serroy, Roman et réalité (677), who is continually disappointed by a novel that seems to promise a modernity that it doesn’t deliver.

70. The alliance of realism with Dutch painting is well known; see especially Yeazell, Art of the Everyday. Less frequently noted, but essential to the eclipse of Heliodorian romance, is the slow decline of the recognition plot; see Cave, Recognitions.

71. Watt, Rise of the Novel, 13–15.

CHAPTER 3. HOW TO READ A MIND (CRÉBILLON)

1. Fielding, Joseph Andrews, 193.

2. Gallagher, “Rise of Fictionality,” 344.

3. Fielding, Joseph Andrews, 193.

4. Lesage, Diable boiteux, 28 (this remark occurs in the 1726 edition); Smollett, Ferdinand Count Fathom, 4. (Crébillon, we will see, uses the same terminology.) Foucault, Order of Things, describes the aim of classification and taxonomy as “determin[ing] the ‘character’ that groups individuals and species into more general units, that distinguishes those units from one another, and that enables them to fit together to form a table in which all individuals and all groups, known or unknown, will have their place” (226). For more specifically literary accounts, see Brooks, Novel of Worldliness, 44–93, and Lynch, Economy of Character, esp. 23–119. Strangely, Brooks and Lynch both regard this conception of character as endemic to the specific culture they are studying; for Brooks, character writing follows from the closed, theatrical nature of a French court life, whereas for Lynch it reflects a British market culture of collecting. Their accounts should be supplemented by the wider perspective of Van Delft, Littérature et anthropologie.

5. “That poetry does aim at generality has long been obvious in the case of comedy, where the poets make up the plot from a series of probable happenings and then give the persons any names they like, instead of writing about particular people as the lampooners did” (Poetics 9; quoted from Russell and Winterbottom, eds., Ancient Literary Criticism, 102).

6. For some examples of French novelists and commentators of these decades who follow Lesage and Crébillon in associating the novel with comedy, see May, Dilemme du roman, 110–16. Admittedly, Gallagher’s purpose is not to credit Fielding personally with the invention of fiction, which simply “appears” in the novel at some point during the years mentioned (“Rise of Fictionality,” 344). She does, however, make clear that the novel of which she speaks was English: it was in England that the cultural promotion of “disbelief, speculation, and credit” (345) led to the advent of fiction. Earlier “components for an understanding of fictionality”—Don Quixote and La Princesse de Clèves are mentioned—“did not gel into either a common knowledge of the concept or a sustained and durable novelistic practice until they coincided in the eighteenth-century English novel” (345).

7. Crébillon’s novel was not translated into English until 1751, but Fielding read French and it was certainly known in the original: Gray wrote that his idea of paradise was “to read eternal new romances of Marivaux and Crébillon,” and it is most likely he was referring to Les Égarements (quoted in Paulson and Lockwood, eds., Henry Fielding, 119).

8. Crébillon, Oeuvres complètes, 2: 71. Subsequent references to the novel will be given parenthetically.

9. Fielding, Joseph Andrews, 194.

10. On the silent world of desire, obscured by the language of decency, see for example Foucault, “Un si cruel savoir,” and Stewart, Masque et la parole, 160–67.

11. Winnett, Terrible Sociability, discusses briefly the example of Crébillon (15–20). On the worldly search for mastery, see especially Brooks, Novel of Worldliness, and Dornier, Discours de maîtrise. A number of critics, however, note that the libertine is in many respects a problematic model for the good reader. I will take up these readings below.

12. “For us as for Meilcour, the drama is first epistemological and then evaluative,” writes Brooks; “we must find out in order to size up, we must move toward the total clarity, definition, and arrest of the portrait” (Novel of Worldliness, 33). This reading would make of the novel both a reflection and an instrument of Norbert Elias’s “civilizing process,” by which the pressures of social constraints transform emotive chivalric man into a self-regulated creature who observes his peers as he carefully controls his affect: “As the behavior and personality structure of the individual changes, so does his manner of considering others. His image of them becomes richer in nuances, freer of spontaneous emotions: it is ‘psychologized’” (Civilizing Process, 399).

13. The affinity between Crébillon’s novel and Enlightenment knowledge production of the kind Foucault describes has been suggested by Dornier, Discours de la maîtrise, 12. Bates, “Cartographic Aberrations,” provides a careful appraisal of the commitment to order on the part of the authors of the Encyclopédie.

14. For consistency, I will refer to the young protagonist as Meilcour, and to the older Meilcour whose memoirs we are reading as the narrator.

15. Bennington, Sententiousness and the Novel, 116.

16. See Force, Molière.

17. For the argument that aristocratic distinction should not be confused with a modern idea of selfhood, see Russo, La Cour et la ville, 139. Bennington, Sententiousness and the Novel, describes Versac’s curiously nonsubversive conception of his own difference or singularity thus: “for Versac, ‘singularité’ does not involve standing outside the closure of worldliness but in becoming such a perfect representative of it in its totality that that perfection itself becomes singular” (118).

18. Pascal’s “Trois discours sur la condition des grands,” written around 1660, was first published in 1670. Versac’s speech contains a number of formulations that recall Pascal’s somewhat functionalist theory of distinction—e.g., “That we must not be inwardly convinced of our merit I will accord—but let it appear that we are” (213).

19. Huxley, Olive Tree, 80.

20. The point is stressed by Bennington, Sententiousness and the Novel, 128. See also Gaubert, “Synchronie et diachronie,” and Garagnon, “Maître à penser.” Contemporaries remarked that Versac’s position was seriously compromised; see the testimony reprinted in Crébillon, Oeuvres complètes, 2: 771, 773.

21. On Crébillon’s portraits, see Jomand-Baudry, “Description.” As Coulet remarks, Marivaux’s portraiture frequently adds singular inflections to basic types (Marivaux romancier, 138).

22. Moreover, both the characters and the narrator use a wide variety of means to undermine or mitigate the apparent absoluteness of their opinions; see Fort, Langage de l’ambiguïté, 135–68.

23. Les Égarements might therefore be viewed not as a novel of worldly discourse, but as a Bakhtinian novelization of worldly discourse. The authoritative text, writes Bakhtin, aims to transmit dead quotations, and as such is profoundly antinovelistic; the novel digests such foreign material, transforming into something new: “One’s own discourse is gradually and slowly wrought out of other words that have been acknowledged and assimilated” (Dialogic Imagination, 345n31).

24. Barthes, “La Bruyère,” 225.

25. Barthes, “La Bruyère,” 230.

26. This and the following quotes from d’Argens can be found in Crébillon, Oeuvres complètes, 2: 769.

27. These judgments, representative of Les Égarements’ reception, are reproduced in Crébillon, Oeuvres complètes, 2: 788 and 790. Writing in the early nineteenth century, Dunlop contends that in Les Égarements “the adventures of more than one individual of rank at the French court are said to be depicted” (History of Fiction, 373), but there is in fact no evidence of readers taking it as a roman à clef—in contradistinction to certain other novels of Crébillon.

28. For contemporary impatience, including letters by Graffigny, see Crébillon, Oeuvres complètes, 2: 770. On the problem of the novel’s completeness, see Coulet, “Les Egarements, roman inachevé?”

29. “Finally, we are not to discover on what moral basis [Meilcour] constructs the values and interpretations which colour his outlook as a mature man” (Adams, “Experience,” 612).

30. Mylne, Eighteenth-Century French Novel, 132; Brooks, Novel of Worldliness, 32–33n6. Brooks’s position, or variants of it, has handily won out over Mylne’s; for a presentation of the critical opinions on the matter, see Edmiston, Hindsight and Insight, 31.

31. See Edmiston, Hindsight and Insight, 33–37.

32. Mylne, Eighteenth-Century French Novel, 132.

33. I take the term “consciousness scene” from Fludernik, Towards a “Natural” Narratology (see esp. 153–59).

34. Haywood, Fortunate Foundlings, 35–36. Haywood’s practice is typical, but I did not choose it completely at random: in 1754, Crébillon published a French adaptation (more than translation) of the work (see Fort, “Heureux Orphelins”).

35. I will take up the subject of unnatural narrators in the Conclusion.

36. Knights’s 1933 essay “How Many Children Had Lady Macbeth?” influential in moving literary criticism from questions of character to themes and structures, derives its persuasiveness largely from its titular question. Yet conflating inquiry into characters’ motives with inquiry into, say, what they had for breakfast is dubious at best. No doubt many literary characters are not designed to elicit psychological scrutiny, but many others manifestly are.

37. Bennington, Sententiousness and the Novel, 122.

38. Crébillon, Oeuvres complètes, 4: 37–38. Crébillon’s theory of illusion in the preface to the Lettres de la duchesse has much in common with Diderot’s work, from the Éloge de Richardson in 1762 to the tales and novels of the 1770s; see Chapter 5.

39. The complexity of Molière’s practice has been explored by Norman, Public Mirror.

40. Contrary to Gallagher’s implication, some of Fielding’s types were (correctly) read as satirical allusions: Peter Pounce, all agreed, was the notorious miser Peter Walter. To discredit his competitor’s inventive gifts, Richardson enumerates with gusto the real-world “originals” of Fielding’s characters (Selected Letters, 197). Of all these authors, no doubt the case of La Bruyère—whose Caractères were routinely printed with marginal “keys”—is the most complex. See Couton, Écritures codées (106–14), and Tourette, “Argument onomastique.”

41. Woloch, One vs. the Many.

42. Cohn, Transparent Minds.

43. See Lynch, Economy of Character. However, in describing a rupture in characterization caused by the developing market culture of late eighteenth-century Britain, Lynch overlooks the many eighteenth-century instances (in France as in England) in which people speak of characters as round, deep, or living. When contemporary talk turns to the excellence of Joseph Andrews, readers appreciate Fielding’s characters—Parson Adams, of course, but also Lady Booby and Slipslop—for their lifelike richness, not for the representativity of their types. (See the testimony gathered in Paulson and Lockwood, eds., Henry Fielding.) Similarly, Brewer, Afterlife of Character, has shown that the lives of characters were routinely prolonged beyond the bounds of a single novel. And Knights’s famous point about Lady Macbeth’s children was directed at a way of thinking about Shakespeare that he traced back, precisely, to the eighteenth century. In other words, the genealogy of the “round” (or deep) character leads through the pseudofactual novel itself.

44. In both Transparent Minds and Distinction of Fiction, Cohn limits her examples of narration of consciousness to Austen and after, but Behn’s use of “consciousness scenes” is treated at length by Fludernik, Towards a “Natural” Narratology (129–77), who somewhat hyperbolically credits the novelist with having provided “an uncanny anticipation of [the novel’s] later perfections” (130). Brown briefly considers the case of Fielding in the context of Walpole’s more extensive narration of consciousness, which he declares the “true source” of what Radcliffe would later perfect (Gothic Text, 19–33, quoted passage on 32). Perhaps some works are indeed sources for later practices: for example, there is some evidence (Neumann, “Free Indirect Discourse”; Bray, “Source”) that Austen’s third-person free indirect discourse may derive rather directly from first-person forms—not Crébillon’s, but Richardson’s epistolary novel Sir Charles Grandison (1753–54). My point is merely that rather than fit individual cases into a preexisting narrative, literary history needs to assess with care the relations between techniques that have at least the potential to be idiosyncratic. For my part, I suspect that the omniscient narration of thoughts is confined to a relatively few, unrelated texts until the turn-of-the-century expansion of third-person fictional (and not pseudofactual) forms of narration that I describe in the Conclusion.

CHAPTER 4. THE AESTHETICS OF SENTIMENT (ROUSSEAU)

An earlier version of this chapter appeared in “Rousseau’s Readers Revisited: The Aesthetics of La Nouvelle Héloise,” Eighteenth-Century Studies 42, 1 (2008), reprinted by permission of Johns Hopkins University Press.

1. Rousseau, Correspondance complète, 10: 104 (henceforth abbreviated CC).

2. On Rousseau’s conflicted relations with Rey over the latter’s fidelity to the author’s intentions, see Turnovsky, Literary Market, 184–203.

3. Rousseau, Oeuvres complètes, 2: 5 (henceforth abbreviated OC).

4. The short first preface, published in the novel’s first edition, was in fact written after the much longer “second” preface. The latter was published separately several weeks after the novel, as Rousseau had long planned.

5. Meltzer, Hot Property, 115.

6. Darnton, “Readers Respond,” 332, 331, 333. Darnton was not the first to propose that Rousseau’s readers believed in his fiction. Bernard Guyon, annotator of Julie for the Gallimard Pléiade edition, assures that “numerous readers believed it to be a true story” (OC 2: 1345n2). But the celebrity of Darnton’s argument has worked the idea into even reference books: “many readers took the letters to be authentic,” reads the Oxford Guide to Literature in French (France, “Julie,” 572).

7. Darnton, “Readers Respond,” 233.

8. Davis, Factual Fictions, 156.

9. Fort, “Accessories of Desire,” 152. Fort references Darnton’s study at this point, as does Vincent-Buffault when she claims that “Rousseau himself prepared the way for this phenomenon of identification since in his preface he presents himself as the mere editor of the letters” (Histoire des larmes, 19).

10. Darnton, “Readers Respond,” 216. In this spirit, Elena Russo divides the eighteenth century into good and bad halves, with an earlier self-aware aesthetic that is in fact more truly modern than a later, naïve one: “disenchanted belief lies at the core of Marivaux’s profound skepticism concerning those virtues that would be extolled by Rousseau and by the Revolution: transparency, sincerity, moral rigorism, and clarity” (Styles of Enlightenment, 261).

11. Thus Séité, Du livre au lire, explicitly opposes the novel to the sentimental readings it was given: “Alongside the legato of effusive and sentimental reading, on which the critical tradition has placed much insistence, Rousseau took care to leave open, indeed to encourage, the possibility of reasoned or reflective reading” (354). In a reply to Darnton, LaCapra has treated the rhetorical complexity of Rousseau’s second preface as a deconstructive rebuttal to the effusive simplicity of his readers’ reactions (“Chartier,” 108–12). In this same vein, and well before Darnton, De Man held that readers mistakenly took the book as “an invitation to a shared erotic or passionate experience” (Allegories, 194).

12. My focus is therefore only on a few key aspects of these letters relating to the nature of the characters and the emotions of reading. For a broader account, see Labrosse, Lire au XVIIIe siècle. Labrosse’s book, which appeared simultaneously with Darnton’s, explores the archive synthetically in view of establishing a typological account of reception, as opposed to the historical one envisioned by both Darnton and myself.

13. The link between the Lettre and Julie was remarked by a number of correspondents, including d’Alembert himself.

14. In this imaginary “conversation about novels,” a reader identified only as “N” asks the author “R” how such an enemy of the theater could possibly stoop to painting La Nouvelle Héloïse’s “vivid situations and passionate feelings”: “Reread the letter on spectacles, reread this collection [of letters]…. Be consistent, or change your principles” (OC 2: 25, ellipsis in original). But the novel’s first preface had already, from its first sentence, obliquely alluded to the Lettre à d’Alembert: “There must be spectacles in big cities, and novels where populations have been corrupted” (5).

15. English novel readers of both sexes, “meditative and solitary, are less given to frivolous imitations” (OC 5: 75); Roussel, “Phénomène de l’identification,” argues for this resolution of the contradiction. Nonetheless, efforts to theorize the difference between spectators and readers remained haphazard throughout the period, notwithstanding occasional squabbles about the effects of images and the effects of words, as when Rousseau argues in the Essai sur l’origine des langues that Du Bos, following Horace, was wrong to say that painting moves us more effectively than poetry (377).

16. Similarly, in the Lettre à d’Alembert, the theater’s existence in big cities can be considered a positive force, since it keeps people out of worse trouble. Whence the paradox, “When the population is corrupted spectacles are good for it, and bad when the population itself is good” (OC 5: 60).

17. Saint-Preux’s program for a type of novel that, like Julie itself, would make people love virtue, has no counterpart in the Lettre (see OC 2: 277). For a treatment of the contradiction between the Lettre and Julie that differs from the one I am offering, see Kelly, Rousseau as Author, 104–15.

18. Hegel, Aesthetics, 1: 1. Or, Kant: “That taste is always barbaric which needs a mixture of charms and emotions in order that there may be satisfaction” (Critique of Judgment, 58).

19. Horace, Satires, 449. Horace’s lines figure prominently in Hénin’s exhumation of the idea of mimetic emotional contagion in the Italian Renaissance (Ut pictura theatrum, 578–96).

20. For an exemplary formulation, see La Mesnardière’s description of the relay by which the theatrical spectator comes to feel “that his heart is like a battleground where a thousand tumultuous passions fight to the death when the poet’s science wills it” (Poétique, 74).

21. On efforts to deal with the paradoxes attending contagion theory, see Hénin, Ut pictura theatrum (597–605), who points out that although the occasional Renaissance commentator allowed for the production of aesthetic emotions qualitatively different from the emotions of the characters, this line of thinking had no real counterpart in subsequent French poetics. Not that neoclassical theorists always spoke in the same terms on the subject. Some, for example, reached for metaphors not of contagion but of “impression.” Hence, Bouhours praises Racine’s tragedies for “a very touching quality that does not fail to impress [on spectators] the passions they represent” (Remarques nouvelles, 93), while Nicole uses the same logic to condemn the theater: “one must not imagine that one can just erase from one’s mind this impression that has been voluntarily excited, nor doubt that it will make us well disposed to the same passion that we so eagerly sought to feel” (Traité de la Comédie, 36–37). Neither contagion nor impression amounts, however, to a psychological theory of affective transmission—what I will be calling identification.

22. Barish, Antitheatrical Prejudice, 276.

23. The novelty of Enlightenment psychological accounts of sympathy with respect to earlier physiologies is stressed by Force, Self-Interest, 28–34; special mention of Malebranche is made on 29. For additional studies of contagion theory as it relates to novel reading, see Toumarkine, “Châteaux en Espagne,” and Fournier, “Pathology of Reading.”

24. Force, Self-Interest, 26.

25. The context is plainly literary, since Crébillon père’s tragedy Atrée et Thyeste (1707) is mentioned. This fragment follows the reasoning of the bulk of the Lettre à d’Alembert: identification or place-switching (se mettre à la place and s’identifier appear to be synonymous) does not occur with villains or cowards (at least in France) because the theater pulls us in through its flattery of the gallant passions that define us as a culture. (The contention, in Poetics 13, that one can feel no fear or pity for bad men appears superficially similar, but Aristotle gives no account of what exactly blocks our fellow-feeling.) Finally, Starobinski, “‘Se mettre à la place,’” offers a look at the concept of identification and place-switching in the period I am covering. In Starobinski’s view, place-switching moves from being considered negatively (in the texts of Nicole and other critics of the theater) to being seen as highly desirable (in Diderot). His account is flawed, however, since contagion was widely spoken of in positive terms by the theater’s proponents; my purpose here is to show there is indeed a shift, but not the one Starobinski describes.

26. The minor distinctions for which Rousseau’s Lettre allows—e.g., spectators do not actually fall in love with the characters, but are instead prepared to fall in love in real life (OC 5: 47)—palliate the obvious absurdities of a model that is never discarded.

27. See Force, Self-Interest, 7–47. For other examinations of Rousseau’s pity in the context of the theater, see Gearhart, Open Boundary, 259–74; and Marshall, Surprising Effects, 144–51.

28. See Viala, Lettre à Rousseau, who approaches many of the issues I examine here, including the relation of Rousseau’s thought on “interest” to what I am calling the proto-aesthetic tradition.

29. Guyon’s comments in his edition of Julie are firmly in this vein (see OC 2: xxii–xxiv); among other examples, see Kavanagh, Writing the Truth, 28.

30. Darnton, “Readers Respond,” 228. For some readings that are particularly attentive to interpretive intricacies of the novel, see Stewart, Half-Told Tales, 115–210; DeJean, Literary Fortifications, 112–90; and Ray, Story and History, 240–69.

31. My estimate of the size of the dossier excludes mere thank-you letters or brief mentions of approval. Remarks on the age, sex, and class of the correspondents can be found in Labrosse, Lire au XVIIIe siècle, 241–44.

32. By far the most emotional response to Lafayette’s work came not from average readers but from the critic Valincour, who devotes roughly a third of his Lettres sur la Princesse de Clèves (1678) to “feelings” stimulated by the novel. In Chapter 1, I have pointed to Valincour’s language of identification, which is notable as much for its difference with respect to dominant metaphors of contagion as for its resemblance to testimony by Rousseau’s readers.

33. Moreover, a good deal of Pamela’s reception was characterized by a somewhat satirical erotic investment: “Rather than being ‘touched’ by Pamela’s narration, actor and audience conspire to do the touching” (Turner, “Novel Panic,” 83).

34. Darnton, “Readers Respond,” 251.

35. Darnton, “Readers Respond,” 245.

36. The small number of other letters from apparent believers (e.g., CC 8: 158, 257) do not fail to articulate their position as a wish: it would be nice if the protagonists existed, because they are such a credit to humanity. And like Du Verger, these readers tend to transform praise of the heroine into adulation for Rousseau’s perfection as a writer.

37. Quoted in Esmein, ed., Poétiques du roman, 140. As I will show in Chapter 6, plenty of writers noted that belief was superfluous to the creation of pleasure in genres such as fairy tales and opera, but this pleasure was held to be distinct from the passions stirred by the highest genres.

38. One of the first to articulate the hypothesis that Rousseau’s original readers believed in the authenticity of the letters, Rétif de La Bretonne, assumes like Roguin that belief only extended to the autobiographical experience underlying the novel: “No one wanted to believe that Jean-Jacques’s Julie and Claire were imaginary beings: everyone said, ‘Jean-Jacques depicted the women he saw and maybe loved’” (quoted in May, Dilemme, 151).

39. A telling example involves d’Urfé’s L’Astrée, which the jurist Olivier Patru took pains to interpret as a “romanced” version of d’Urfé’s own affair with Diane de Châteaumorand. Patru’s argument, however, did not involve the irreplaceability of originary authorial emotion, but treated d’Urfé’s ingenious use of his own experience as an analog to the dramatist’s use of historical events. Through his art, Patru wrote, d’Urfé adds interest to a story that otherwise would have none (“Éclaircissements,” 559).

40. The author recognized in his Confessions that this assumption lay behind his female readers’ interest in him: “Everyone was persuaded that one could not express so vividly feelings that had not been lived, nor paint love’s transport, if not from one’s own heart” (OC 1: 548).

41. Diderot’s Éloge de Richardson can be seen to operate a similar revision in theories of identification; see Chartier, “Richardson,” esp. 663–64. The idea of quixotic identification was still current in Rousseau’s time, both in philosophy—“There is no man, I would hazard, who in his moments of boredom does not imagine some novel of which he is the hero” (Condillac, Essai, 122–23)—and in poetic treatises: “You [the writer] transform a weak heart into a magnanimous one, and the members of your audience, at least for a time, become so many heroes,” writes Rémond de Sainte-Albine in 1747 (quoted in Hobson, Object of Art, 182). From this point of view, the mid-century interest in absorption exemplified by Diderot may be seen less as a reaction against rococo theatricality, as Fried has famously claimed (Absorption), than as a displacement of older theories of identification.

42. In the Lettre, by contrast, Rousseau had followed Du Bos and Nicole in claiming that “a man without passions, or who could completely control them, could never move [intéresser] anyone” (OC 5: 17).

43. “[Women] everywhere have read your Héloïse, but they haven’t budged from the study and the 55th letter of the first book,” writes one critic, mentioning the letter that marks the apogee of Saint-Preux’s requited passion for Julie (CC 8: 316). The possibility of a reader who fails to complete the entire Rousseauian “cure” becomes a plot device in a number of turn-of-the-century English novels, such as Opie’s Adeline Mowbray (1805), in which the heroines suffer because they are not allowed to finish La Nouvelle Héloïse; these female protagonists undergo the contagious emotional influence of adulterous love without ever learning the good news of Julie’s renunciation. See Grogan, “Politics of Seduction.”

44. Vila, Enlightenment and Pathology, 199–200. To be sure, the reading is sanctioned by Rousseau himself in the second preface, who makes use of the metaphor via a quote from Tasso (OC 2:17).

45. Lucretius’s proposition, much discussed by Du Bos, and echoed by Hume and Burke, reads: “It is comforting, when winds are whipping up the waters of the vast sea, to watch from land the severe trials of another person” (Lucretius, On the Nature of Things, 35; on Du Bos’s discussion, see Marshall, Surprising Effects, 22–25). Lucretius goes on to explain such pleasure in the customary Epicurean manner—that is, as an effect of the observer’s relief at being safe from danger. Rousseau’s rewritings, it should be noted, entirely avoid any recourse to self-interest.

46. Burke’s A Philosophical Enquiry into the Origin of the Our Ideas of the Sublime and Beautiful dates from 1757. Kant’s Observations on the Feeling of the Beautiful and the Sublime, which cites Rousseau (though not Julie), is from 1764.

47. “This meeting of two hearts that cry in solitude at the same time is a summit, a suspension of subjective transparency” (Vincent-Buffault, Histoire des larmes, 48).

48. Thus, Ray, “Reading Women,” has argued that the second half of the novel deliberately puts into place a “cognitive mode of pleasure quite unlike the erotic suspense and sentimental extravagance of the earlier episodes” (430). Ray’s reading is confirmed by the testimony of Rousseau’s contemporaries that we have already seen.

49. De Man, Allegories, 215.

50. “La Nouvelle Héloïse would be a very different (and a much shorter) text … if the narrative had been allowed to stabilize in this way” (De Man, Allegories, 215).

51. Significantly, what Claire cannot yet do, Wolmar points out, is cry (OC 2: 740). Though we might think of this as prefiguring twentieth-century psychologies of grief, it is more in keeping with the thrust of the novel to understand Claire’s inability to cry in proto-aesthetic terms. Crying is a critical acknowledgment of distance, and not the mark of total loss of self in the spectacle that we often associate with the sentimental reader.

52. “But who still cries at the theater?” asked Roland Barthes, alluding to the reception of Bérénice (Sur Racine, 142). His call for a history of crying (and the eclipse of aesthetic tears) has been at least partially told by scholars such as Vincent-Buffault, Histoire des larmes, and Elkins, Pictures and Tears.

53. My reading of the second preface therefore conflicts with de Man’s, who sees its concerns with referentiality as a signal displacement of older debates on history and verisimilitude (Allegories, 195–96). But Rousseau does not leave such debates in the dust of what de Man praises as “a new sense of textuality” (204); he merely approaches them from an unusual angle.

54. Goldhammer, “Man in the Mirror,” 42.

55. Sermain, Singe de don Quichotte, extensively examines the importance of the Quixote for the eighteenth-century novel’s self-definition. For a slightly different take on Rousseau’s use of Cervantes, see also Sermain, “La Nouvelle Héloïse,” 234–36.

56. “People didn’t wait for Rousseau to cry while reading,” notes Vincent-Buffault, Histoire des larmes, 14. On pathétique specifically, see Coudreuse, Goût des larmes. Appropriately, given Rousseau’s attention to the play in the Lettre à d’Alembert, Bérénice stands as the key reflection on how tragedy might be geared to the sole production of emotion. For many reasons, Julie can fruitfully be considered an elaborate amplification of the tragedy that notoriously turned its back on elevated subject matter and the intricacies of plot in favor of the production of tears alone. Various examples of early modern tears in France are explored in Cron and Lignereaux, eds., Langage des larmes, while Ibbett, “Pity,” examines the evolution of cathartic vocabulary.

57. Wordsworth, Preface to Lyrical Ballads, in William Wordsworth, ed. Gill, 611, 599.

58. This is my position in the article on which the present chapter is based, “Rousseau’s Readers Revisited.”

59. Figures on editions come from Rousseau, Nouvelle Héloïse, ed. Mornet, 1: 238. According to Mornet’s figures, even a book as influential as this did not have the sweeping effect on production that one might think: only about a fifteenth of novels published between 1761 and 1789 are obviously Rousseauian (1: 291).

CHAPTER 5. THE DEMON OF REALITY (DIDEROT)

1. Champfleury, Réalisme, 90 (following quote, 273).

2. Diderot, Contes et romans, 449. Unless otherwise stated, all quotations from Diderot’s prose narratives will be taken from this edition (henceforth abbreviated D). For Barthes’s description of the reality effect, see “L’Effet de réel.”

3. Champfleury is not without followers, however. For an argument that Diderot’s details signal a proto-realist approach to versimilitude, see Joly, Deux Études.

4. Undank, “Jacques le fataliste,” 752. Diderot is thus assumed to discover a modern truth about literature that is at the furthest possible remove from “naïve” sentimentality: “If, after a tentative effort at a modified version of the Richardsonian mode in La Religieuse Diderot moved to the opposite extreme, flaunting artifice and eschewing the minute circumstantiality he had extolled in Richardson, it was not only because of his progressive disenchantment with ‘sensibility’ during the later 1760s and early 1770s but also, I believe, because he came to see more clearly that no fiction could really escape from its conspicuous condition as an arbitrary ordering” (Alter, Partial Magic, 65). I will return to other such readings presently.

5. The Correspondance littéraire, philospohique et critique was supported by subscriptions from various European courts, including that of Catherine the Great. It appeared from 1753 to 1793, and included, along with accounts of the cultural goings-on in Paris, numerous texts by Diderot. (The diffusion in the Correspondance littéraire can be called publication, since the early modern sense of the word was a rendering public, rather than an actual printing.) The view of the two novels as counterparts is expressed in a letter of September 1780 (Diderot, Correspondance, 15: 190).

6. The term “Préface-annexe” was first used by Jules Assézat in his 1875–77 edition of Diderot’s Oeuvres complètes. (Diderot’s novels are found in volume 5.)

7. I quote Grimm’s original account from volume 7 of Naigeon’s 1798–1800 edition of Diderot’s Oeuvres (here, 267; henceforth abbreviated G).

8. In a sense, the memoir has never stopped being fragmentary: critics have repeatedly pointed out various oddities and contradictions, of which the most persistent is a shifting temporal point of view. For a quick account of some of these contradictions, see May, Diderot et “La Religieuse", 206–8; more forensic detail can be found in Parrish, “Conception,” 361–84. Moreover, the story of the hoax itself is riddled with inconsistencies, possibly due to the ten-year gap between the original events and Grimm’s recollection of them; see Mylne, “Truth and Illusion.”

9. Booy, “Inventaire,” 392.

10. In a letter to Sophie Volland dated September 17, 1761, Diderot tells of breaking off his reading of Clarissa in order to address his lamentations to Richardson’s characters; a friend, d’Amilaville (not yet d’Alainville) “couldn’t make head nor tail of my transport or my words” (Correspondance, 3: 306). In the Éloge de Richardson, Diderot claims to have observed this behavior in an unnamed friend (D 908). In both cases, the self-consciousness of the creator who makes himself cry is not yet present.

11. Buisson, the first publisher of La Religieuse, seems to have had access to Grimm’s library, which had been confiscated in 1792 due to its owner’s emigration. But the configuration of his 1796 edition is a mystery. Were there, in Grimm’s papers, copies of the manuscript of the memoir and the original hoax account of 1770 but not the modified version that appeared in the March 1782 issue of the Correspondance littéraire? If so, was it merely coincidence that Buisson decided to position Grimm’s text after and not before the memoir, thus unwittingly fulfilling Diderot’s apparent intention on that score? As of yet, we have no answers to these questions. For information on the provenance of the Buisson manuscript, see Booy and Freer, eds., Jacques le fataliste et La Religieuse devant la critique, 11–36.

12. Dieckmann expresses understandable perplexity at the fact that Assézat did not see this title (“Préface-Annexe,” 38–39n7).

13. Dieckmann, “Préface-Annexe,” 31. Curiously, Dieckmann was as unaware as anyone else that the original appearance of La Religieuse in the Correspondance littéraire had been followed by the expanded preface, and therefore still needed to mount an argument for its inclusion and postpositioning.

14. Dieckmann, “Préface-Annexe,” 22–23.

15. Diderot, La Religieuse, ed. Mauzi, xi.

16. Varloot, “La Religieuse et sa Préface,” 268, 269.

17. Although the argument of Success in Circuit Lies appears to be influenced by a famous series of poststructuralist readings of Poe’s “Purloined Letter” (see Muller and Richardson, eds., Purloined Poe), De la Carrera may also be following a hint by Mylne to the effect that the preface’s meditation on truth and illusion can be extended to the events of the hoax itself (“Truth and Illusion,” 356). (De la Carrera herself mentions neither.)

18. Frigerio, “Nécessité romanesque,” 54. Frigerio references De la Carrera repeatedly.

19. Russo, Styles of Enlightenment, 23. For more such studies, see Mortimer, “Naïve and Devious,” and Terdiman, Body and Story, 19–38. Hobson, Object of Art, contains a number of references to the Préface-annexe and its “hyperconsciousness” (178).

20. Isolated remarks in longer studies do begin to take the Préface-annexe in less ironic directions. See Kempf, Diderot et le roman, 221; Stewart, Imitation and Illusion, 308; and Mylne, Eighteenth-Century French Novel, 212–13. None of these scholars offer, however, a full-scale reading of the Préface-annexe.

21. Lafon, in D 1264; Bryson, Word and Image, 198–99.

22. Varloot, “La Religieuse et sa Préface,” 270. Mauzi puts it thus: “With the Préface-annexe, La Religieuse is already, as M. Dieckmann suggests, Jacques le fataliste” (Diderot, La Religieuse, xi). For Dieckmann’s original remark, see Préface-annexe, 29.

23. The modern critical need to save sentimental texts from their sentimentality is not restricted to the work of Diderot: other eighteenth-century novels have gone through a process of revisionism by which former heroes are reinterpreted as hypocrites and fools. See for example Nünning, “Unreliable Narration.” La Religieuse is unusual only to the extent that the Préface-annexe suggests (or so the argument goes) Diderot’s own supposed evolution on the issue.

24. Russo, Styles of Enlightenment, 23.

25. It should be noted that attributions of the original 1770 hoax account to Diderot instead of Grimm are perilously speculative. The elaborate allegory of mastery developed by De la Carrera—according to which the entire hoax may well be a hoax played on the reader—is predicated on just such a “hunch” (Success in Circuit Lies, 217n14). Though De la Carrera admits that her hunch is not supported by extant documentation, she implies that Grimm’s authorship is equally unverifiable. (Terdiman advances a similar argument, apparently without knowledge of De la Carrera, in Body and Story, 30–32.) This is misleading, however: Grimm’s voice is that of the Correspondance littéraire; we have no examples of Diderot’s—or anyone else’s—tacit usurpation of that voice; and at any rate, the various changes Diderot made to the original account—especially his aggressive reaction to Grimm’s assertion that the memoir had never been finished because of Diderot’s work habits—make little sense if we assume that Diderot had authored the first version.

26. Caplan, Framed Narratives, 45, 79.

27. Were it not to his liking, Diderot could have of course suppressed Grimm’s mention of the conspirators’ laughter. Two comments are in order. First, for reasons unknown, Diderot changed without cutting Grimm’s text: he makes himself the hero of the hoax by replacing “we” with his own name; but he leaves in place Grimm’s jab at his work habits, merely defending himself in an aside. Second, Diderot—like many others—clearly loved jokes and put-ons; see for example Kempf, Diderot et le roman, 211–32; Catrysse, Diderot et la mystification, 21–59; and, for the broader context, Bourguinat, Siècle du persiflage, esp. 82–97. Joking, then, was in the air; but if jokes can have a literary lesson, it is not the lesson critics have taken to attributing to the joke played on Croismare.

28. Mirzoza supposes that the African will not actually be taken in, but only because current theatrical practice is so artificial; see D 136–37. For the Discours’ comparison of the ideal spectator and the child, see Diderot, Oeuvres esthétiques, 215.

29. Jaucourt, “Opéra,” 495.

30. Mercier, Du théâtre, 104n.

31. The anecdote is reported in Roubine, “L’Illusion et l’éblouissement,” 410–11.

32. See, for example, d’Aubignac’s admiring reference to a girl who so pities Pyramus deploring the supposed death of Thisbe that she wants her mother to warn him that his lover is in fact still alive (Pratique du théâtre, 427 [IV, 7]). The Zeuxis anecdote is found in Book 35 of Pliny’s Natural History.

33. Though anecdotes attesting to perfect aesthetic immersion are hardly new, as I’ve said, Mercier’s is probably a response to Rousseau’s contention that the (reprehensible) pleasure of theater came from the fact that spectators could enjoy the sufferings of others because they were absolved of the obligation to do anything to help them. (See Rousseau’s Lettre à d’Alembert, in Oeuvres complètes, 5: 23–24.) Diderot’s novel should be seen as part of this polemic as well.

34. As Marshall notes in Surprising Effects of Sympathy, 85. In Marshall’s opinion, the Préface-annexe cannot help tainting the good-faith sympathy of Croismare with “anxious reflections about narrative truth and artifice” (85), and he extends this lesson to Simonin’s entire memoir, which thematizes the dangers associated with “the contagious disease of self-forgetting or self-loss called sympathy” (103). Marshall’s argument is deconstructive, since the good or pure quantity is shown to be constitutionally inseparable from its devalued antithesis: sympathy is always already fraud; sincerity is never anything but theatrical. Without pretending to offer a full reading of La Religieuse, I can at least suggest that Diderot’s attention to sympathy’s potentially dangerous or lubricious variants does not undermine his commitment to it, on the contrary.

35. Beaumarchais, Oeuvres, 139. Though Beaumarchais was indebted to Diderot for the latter’s pioneering theorization of the drame, dates suggest that the “Question for Men of Letters” must be an expansion of Beaumarchais’s remarks.

36. Booy and Freer, eds., Jacques le fataliste et La Religieuse devant la critique, 157 (henceforth abbreviated BF). Here and elsewhere, “fiction” has the customary meaning of a lie or untruth. Devaines’s remark on La Religieuse resembles the argumentation Richardson put forth to object to Warburton’s “outing” of him as Clarissa’s author; see the Introduction.

37. Apart from Devaines and Naigeon, only one other reviewer expresses dismay at the “disenchant[ing]” effect of the Croismare anecdote (BF 294).

38. Assézat, publishing the Préface-annexe, noted as much; see Diderot, Oeuvres complètes, 5:205. See as well the observation of Coulet: “Naigeon’s protests rang hollow: La Religieuse was not presented as a true story, at least not any more than any other memoir novel” (Roman jusqu’à la Révolution, 465). One could argue that its truth pretense was actually weaker than that of typical novels: it had no preface at all attesting to the origins of the found manuscript.

39. In a previous review, Duval praises the quality of the hoaxers’ writing by saying, “I’m shamed to admit that, as aware as I am of the falseness of these letters [between Croismare and Delamarre], once I start reading one I can’t get past the middle without my eyes watering and my heart constricting” (BF 124–25).

40. Roederer and Duval are the only reviewers to make the case that the Grimm text does not change the reader’s pleasure, just as Devaines and Naigeon are nearly the only ones to deplore its inclusion. Most take no notice at all of the appended text.

41. “Le procédé de la verrue,” writes May (Quatre Visages, 203). It is frequently pointed out that Diderot seems to be repeating a contention that appears in La Nouvelle Héloïse, where Saint-Preux criticizes a portrait of Julie for leaving out her defects, including “the little scar left under your lip” (Oeuvres complètes, 2: 291). The Rousseau passage, however, speaks about accuracy and not illusion.

42. Although most have taken the theory of the detail in Les Deux Amis at face value, two scholars have arrived at readings that are more or less explicitly influenced by previous treatments of La Religieuse. Jean Ehrard argues that Diderot actually awakens the good reader from illusion by inserting in the text various factual contradictions, thus making “truth … problematic” (Invention littéraire, 161). For Pierre Chartier, Diderot’s remarks on illusion at the end of Les Deux Amis warn us that in spite of the uncanny fusion of reality and fiction, the text is a text, and that “all realism is mystifying” (“Le Conte ‘historique,’” 11); Diderot thus manages simultaneously to set the realist nineteenth century on its way and to “initiat[e] the anonymous readership of novels, at least virtually, into the modern era of suspicion.” Both readings are built not on Diderot’s text and context, but on the familiar poststructuralist belief that realism is naïve.

43. Diderot, Correspondance, 10: 124. A cursory reading of Grimm’s literary criticism in the Correspondance littéraire confirms that his tastes were considerably less sentimental than Diderot’s.

44. Statistics can be found in Martin, “Origins of the Contes moraux” and “Marmontel’s Successors.”

45. Diderot, Correspondance, 10: 124.

46. Decades later, when Marmontel was writing the new tales of a more distinctly sentimental nature that appeared from 1789 to 1792 in the Mercure de France, he took care to add a variety of pseudofactual frames.

47. Saint-Lambert, Contes, viii–ix.

48. The names of Scarron and Cervantes—whose tales were rich in murder, vengeance, duels, abductions, and so on—hint that Diderot is thinking back to a time when prose narrative did aspire to high seriousness. (Likewise, Rosset’s roughly contemporaneous Histoires tragiques [1614] was a widely reprinted collection of various tales of criminal passion designed for the edification of its readers.)

49. On the naturalizing of eloquence in the second half of the seventeenth century, see Declercq, “Rhétorique classique.”

50. See, e.g., Reiss, Toward Dramatic Illusion.

51. Mylne, Eighteenth-Century French Novel, 195.

52. Mylne, Eighteenth-Century French Novel, 198.

53. The text—we cannot know how closely it conforms to what was read by Naigeon—can be found in D 451–62; for information on the discovery, see Varloot, “Une version originale.”

54. Bongie, Diderot’s femme savante, 93–94.

55. In Saint-Lambert, “material life is weightless,” astutely observes one critic, whereas Diderot’s heroes “are crushed by the burden of their poverty” (Ehrard, Invention littéraire, 155).

56. The peculiar quality of Félix’s stammering—“Kill me, kill me” does not so much “express” the plenitude of the speaker’s pain as it allows us to glimpse that pain under rhetorical poverty—may well come straight from Diderot’s work on the theater, where true emotion was said to break down the character’s ability to express himself. “In the spectacle of a man animated by some great passion, what affects us?” asks Diderot in the Discours sur la poésie dramatique. “Is it his ability to speak well [Sont-ce ses discours]? Sometimes. But what never fails to move is cries, barely articulated words, broken voices, a few monosyllables escaping here and there” (Oeuvres esthétiques, 101).

57. Baculard d’Arnauld, Délassements de l’homme sensible, 1: 30.

58. La Fontaine, Amours de Psyché et de Cupidon, 123 (La Fontaine alludes to Plato’s Ion [356a]); Horace, Satires, 449 (“If you would have me weep, you must first feel grief yourself”). In such a context, Diderot’s celebrated d’Alainville anecdote is merely a reactivation of the kind of old commonplace one finds in, say, La Mesnardière’s description of the poet weeping over the fate of his characters (La Poétique, 73).

59. Diderot, Oeuvres, ed. Naigeon, 12: 346n. Indeed, phrases like “this tale, which is not one” occurred with regularity in the output of the time (Martin, “Marmontel’s Successors,” 229).

60. Reading Diderot through the lens of the modern artistic practice of putting titles and artworks in dialectic tension or of creating art through the mere act of titling strikes me as too patently anachronistic to require treatment here. Anachronism does not preclude interest, however: see Pucci, “Negative Framing,” and Fleming, “Ceci n’est pas un conte/Ceci n’est pas une pipe.” (Fleming points out that it is almost certain that Magritte was consciously appropriating Diderot’s title.)

61. Bongie, Diderot’s femme savante, 11.

62. In both the original version appearing in the Correspondance littéraire and Naigeon’s first publication, the title does not in fact precede the text: it is put in the mouth of the narrator-author, but set off from the rest of the text (centered, and preceded and followed by skipped lines). It is therefore possible to argue that the title is not quite a title at all.

63. A further complication should be noted, even if space concerns do not allow me to pursue its implications. Ceci was part of a series of dialogued works appearing in the Correspondance littéraire, including Madame de La Carlière (also a tale of imperfect love purportedly drawn from Diderot’s experience) and the Supplément au voyage de Bougainville (in which the speakers read and comment on speeches by and interviews with Tahitians, which were said to have been censored from Bougainville’s Voyage autour du monde). All three texts worked together, as Grimm hinted when presenting the first, Ceci n’est pas un conte: “We shall see only at the end of the last the moral and the secret goal [Diderot] has set himself” (quoted in D 1079n4). In brief, Diderot’s moral point is that the social channeling of “natural appetites” is responsible for the suffering experienced by the French characters of Ceci and Madame de La Carlière (Supplément, D 580).

64. I take all these details from Bongie, “Retour à Mademoiselle de la Chaux.”

65. See Bongie, “Retour à Mademoiselle de la Chaux.”

66. See Bongie, Diderot’s femme savante, 85.

67. Pucci, in “Negative Framing,” has advanced a reading that does goes in the direction of prevalent treatments of the Préface-annexe. Her argument, situated on an abstract and theoretical level, takes no note of Diderot’s evident concern with the real-world referents of his characters.

68. See, e.g., Diderot, Quatre Contes, ed. Proust, xli. Similarly, some critics await the identification of historical referents of other memorable protagonists of Diderot’s novelistic works, notably those of the celebrated tale of a mistress’s revenge in Jacques le fataliste: “The day will come when we will know [the identities of] Madame de La Pommeraye, le Marquis des Arcis,” is the optimistic opinion, a half-century old now, of a scholar who assiduously chased down the origins of many of the novel’s numerous anecdotes (Vernière, “Diderot et l’invention,” 164). In Diderotian criticism there is something of a running debate on Diderot’s capacity for invention (perhaps all his novels, some say, are never anything but reworkings of anecdotes); see the remarks of Bongie, Diderot’s femme savante, 30–34.

69. Bongie’s sleuthing is exemplary, but no more than Naigeon does he even begin to explore the interpretive difficulties presented by the fact that Ceci n’est pas un conte is in fact one part tale and one part not-a-tale. To my knowledge, no one has.

70. May, Quatre Visages, 164–65.

71. Diderot builds an additional level into his meditation on truth and reference by having La Chaux author Les Trois Favorites, a “little historical novel” which may—or may not—allude to the Marquise de Pompadour. While space concerns preclude treatment here, I can at least point out that if Ceci n’est pas un conte is unusual in its insistence that a “true story” might actually be true, Les Trois Favorites works the opposite way by insinuating that a text having all the marks of being a gossipy roman à clef may in fact be just what it says it is—a flight of fancy.

72. Bongie, Diderot’s femme savante, 38, 89, 34, 21.

73. Bongie, “Retour,” 103.

74. Bongie, Diderot’s femme savante, 139, 99, 36.

75. Champfleury, Réalisme, 95.

CHAPTER 6. BEYOND BELIEF (CAZOTTE)

An earlier version of portions of this chapter appeared in “Permanent Re-Enchantments: On Some Literary Uses of the Supernatural from Early Empiricism to Modern Aesthetics,” in The Re-Enchantment of the World: Secular Magic in a Rational Age, ed. Joshua Landy and Michael Saler (Stanford, Calif.: Stanford University Press, 2009), © 2009 Board of Trustees of the Leland Stanford Jr. University, reprinted by permission.

1. Freud, “Uncanny,” 368; further passages quoted in this paragraph are from 405, 400, 389.

2. Most specialists of the fantastic define the genre along the lines laid down by Freud’s definition of the uncanny, carefully separating it from fairy tales and the like. In the words of the author of the classic work on the French fantastic, it is constituted by “the brutal intrusion of mystery into the domain of real life” (Castex, Conte fantastique, 8). The much-cited analysis of Todorov, Fantastic, which I will take up presently, seconds the nineteenth-century location of the genre. A comparative overview can be found in Bessière, Récit fantastique. For an exploration of the gulf between fairy tales and properly fantastic narratives, see Caillois, “De la féerie.”

3. Freud, “Uncanny,” 401–2.

4. Clery, Rise of Supernatural Fiction, esp. 13–32.

5. Castle, Female Thermometer, 162. Castle’s many earlier examples of the uncanny are extraliterary.

6. Nelson, Secret Life, 18. This is a commonplace. Caillois: the fantastic “is born at the moment when everyone is more or less persuaded of the impossibility of miracles,” and serves “as compensation for an excess of rationalization” (“De la féerie,” 19, 27). Rosemary Jackson: “It is hardly surprising that the fantastic comes into its own in the nineteenth century, at precisely that juncture when a supernatural ‘economy’ of ideas was slowly giving way to a natural one” (Fantasy, 24). Tobin Siebers: “The Romantics associated superstition and poetry because their distance from magical thought allowed them to transform it into an aesthetic formula” (Romantic Fantastic, 21). The idea may come from the pioneering work of Varma, Gothic Flame, 206–31.

7. Thus one noted critic, analyzing the work of Wilde and James, finds full-blown skepticism in place only at the end of the nineteenth century, just where he needs it; see Ziolkowski, Disenchanted Images, 230. One might remark that Caillois and Jackson, quoted in the previous note, also sprinkle their affirmations with “preciselys” and “more or lesses.”

8. “Quodcumque ostendis mihi sic, incredulus odi”: “Whatever you thus show me, I discredit and abhor,” writes Horace (Ars poetica, in Satires, 466–67). The poet is in fact referring not to anything supernatural, but to actions that revolt reason and morality—Medea’s murder of her children and the dinner Atreus makes of his brother Thyestes. Nonetheless, in the eighteenth century incredulus odi was routinely directed against things mythological or supernatural.

9. The standard account of the period’s refusal of the pagan marvelous can be found in Bray, Formation, 231–39. On Médée as neoclassicism’s other, see, for example, Greenberg, “Mythifying Matrix”; Greenberg’s work, here and elsewhere, gives a Lacanian twist to the well-known arguments of Elias, Civilizing Process, and Weber, “Science as a Vocation.” Accounts like these often take this development as a sign of deep change, reflecting a key stage of rationalism (Bray) or the evolution of modern subjectivity (Greenberg).

10. On the early modern use of the Medea myth, see Wygant, Medea; Corneille’s contribution is analyzed on 103–25.

11. Samuel Johnson, Lives of the Poets (1779), quoted in Wellek, History, 81–82.

12. Naturally, the history of the literary marvelous is more complicated than this brief treatment can describe. For the Renaissance material, see Hathaway, Marvels and Common-places. Bray’s presentation of the seventeenth century in Formation should be supplemented by Forestier, Passions tragiques (161–90), who examines tragedy’s use of purely “poetic” marvels. The long erosion of the prestige of marvelous recognition plots is covered by Cave, Recognitions. On the constitution of tragedy and opera as part of an oppositional generic system, see Kintzler, Poétique.

13. Chapter 4 contains a more detailed account of passionate contagion.

14. The words appear in the “Argument” of Andromède (1651); see Corneille, Oeuvres complètes, 2: 448. Corneille’s “for the eyes alone” implies a hierarchy of worthiness that is evident elsewhere: “In the end,” writes Chappuzeau in 1674 of another Corneille machine play, Le Toison d’or (1661), “these pretty spectacles are only for the eyes and ears, they do not touch the depths of the soul; and when you return home, you can say that you saw and listened, but not that you were instructed” (Théâtre français, 48).

15. Quoted in Kintzler, Poétique, 254; assembling remarks like these, Kintzler offers a thorough reconstruction of the theoretical separation of tragedy and opera in the period. Corneille’s “for the eyes alone” is often taken as evidence of neoclassical tragedy’s essentially verbal prejudice, of its qualitative separation from, say, Diderot’s later theories of the tableau (see, e.g., Frantz, Esthétique, 12–19). In fact, as I point out in “Fourth Wall,” regular tragedy is commonly discussed as a visual illusion. What is important to recognize is that though tragedy and opera both operated through the eyes, the nature of that operation differed.

16. Robert, Conte de fées littéraire, 171. Robert’s thesis provides a thorough overview of French fairy tale production.

17. The link between fairy tales and the Modern camp is well known. For a particularly informative account of the genre as metafictional reflection, see Sermain, Métafictions, where the phrase “manifesto of modernity” occurs on 358. The genre’s implicit self-positioning vis-à-vis classical rhetoric is explored by Fumaroli, “Contes de Perrault.” Work on women’s authorial involvement in the fairy tale is extensive; see, e.g., Seifert, Fairy Tales, Sexuality, and Gender, and Hannon, Fabulous Identities.

18. Addison, “Pleasures of the Imagination” (1712), Spectator, 3: 570.

19. Bernard, Inès de Courdoue, 80; Durand, Petits soupers, 158.

20. Sermain, Métafictions, 393–400.

21. Addison, Spectator, 3: 570.

22. Stewart, “Notes on Distressed Genres,” 74. Stewart’s account is valuable, though it insists too heavily on the “mournful” and ideologically retrograde character of the distressed genre; many of her remarks on the French context are also factually misleading.

23. Presenting his translation, Galland notes that if tales in general divert because of their marvelous content, “these [tales] beat all others that have yet appeared, since they are full of details that surprise and grip the mind, and that illustrate the extent to which the Arabs surpass all other nations when it comes to such compositions” (Mille et Une Nuits, 1: 21). In associating the Orient with imaginative prowess, Galland was following earlier scholars such as Huet, who credited the Arabs with mastering “the art of lying pleasurably” (Origine, 453). On Galland’s importance to metafictional reflection at this time, see Sermain, Métafictions, 400–410.

24. May, Mille et Une Nuits d’Antoine Galland. May provides a detailed look at why a work as foundational and successful as Galland’s no longer fulfills our criteria of readability.

25. Hurd, Letters, 94–95.

26. E.g., Clery, Rise of Supernatural Fiction, 17–18.

27. Addison, Spectator, 3: 572.

28. The profound structural similarity between the aesthetic landscape of France and England is overlooked, for example, in Clery and Miles’s valuable compendium, Gothic Documents: the purported Britishness of the antineoclassical sensibility is cemented by excluding the types of earlier reflection I have been canvassing. Patently, a novel such as Lhéritier de Villandon’s La Tour ténébreuse (said to be from an English chronicle!) not only belongs in the gothic lineage, but illustrates the genre’s genealogical proximity to the fairy tale (of which Lhéritier wrote not a few). Doubtless, there are cultural reasons for the fact that a properly gothic literature did not develop in eighteenth-century France, and for the converse fact that fairy and exotic tales were the French’s preferred form of fancy; my point is merely that the absence of a “French gothic” has nothing to do with the country’s purported affinity with reason and vraisemblance. (Brown, Gothic Text, has recently argued for the gothic as a pan-European phenomenon; but his comparative study, to which I will return below, does not consider Continental interest before the turn of the nineteenth century.)

29. On the French opéra comique, which rejected mythology in favor of spectacles the sentimental viewers could identify with, see Thomas, Aesthetics of Opera, 179–292.

30. Cazotte, Diable amoureux, 13; further references to the novel will be given parenthetically in the text. Incidentally, Cazotte never explains why the devil speaks Italian.

31. For more on Gabalis, still well known in Cazotte’s time and a key text in the history of literary transformations of superstitious utterances into entertainment, see Sellier, “Invention d’un merveilleux,” and especially Sermain, Métafictions, 137–59.

32. Subligny’s ghost stories are but one of a string of texts in which some characters debunk the beliefs of others. These range from the plays La Devineresse (Th. Corneille and Donneau de Visé, 1680) and La Comète (Fontenelle, 1681) to a string of turn-of-the-century novels—d’Aulnoy’s Relation du voyage d’Espagne (1691), Durand’s Comtesse de Mortane (1699), and Murat’s Lutins de château de Kernosy (1710). (Some of these examples, and others, are discussed by Sermain, Métafictions, 341–53.)

33. Cuillé, Narrative Interludes, has isolated the novel’s numerous references to opera, arguing on the basis of these that its marvelous owes much more to the lyric stage than to fairy tales (71–88).

34. The frame narrative of Caylus’s Contes orientaux (1743) claims similar narcotic properties for its contents.

35. This is the revised ending of the novel, which appeared in the second edition of 1776. In this edition’s epilogue, Cazotte reports the dissatisfaction of his friends on reading the 1772 ending, held to be too “brusque” in its conclusion (88): there, on the ride home, Alvare suddenly realizes the danger and summons the devil to withdraw, at which point Biondetta vanishes and the huge camel’s head appears in clouds that quickly disperse in the wind. He continues on to his mother’s; let this be a lesson to him, she admonishes.

36. Castex, Conte fantastique, 25. Cazotte’s influence on the German fantastic is explored by Von Mücke, Seduction of the Occult, 6–7, 35–57.

37. Todorov, Fantastic, 41.

38. Todorov, Fantastic, 44. One of the problems with Todorov’s grid (I will point out others presently) is that its symmetry is faulty: one can see how a “marvelous” narrative differs from a “fantastic-marvelous” narrative, since the former simply presents marvels as real (no hesitation); but it would seem that there can be no “uncanny” narratives, only “fantastic-uncanny” narratives, since by definition the uncanny is a strangeness that causes us to hesitate.

39. Todorov, Fantastic, 24–25.

40. Todorov, Fantastic, 166.

41. Todorov, Fantastic, 24.

42. Todorov, Fantastic, 85.

43. Todorov, Fantastic, 25.

44. Todorov, Fantastic, 27.

45. Stewart, Rereadings, 210, 206.

46. On these points, see Hoffmann, “Ruse du diable,” and Schuerewegen, “Pragmatique et fantastique.”

47. For this reason, pressed to fit the text into Todorov’s grid, we would have to put it in the “marvelous” category, not in the “fantastic-marvelous” category—in which case, of course, it should not figure in Todorov’s book any more than “Snow White.” At least one critic has tried to argue that the text’s irresolvable contradictions make it an example of the pure fantastic; see Cardinal, “Diable amoureux,” and, much more briefly, Hunting, “Mille et Une Sources,” 270n32. This may, however, be to confuse contradiction and ambiguity: with “pure” fantastic hesitation, we do not have enough information to conclude one way or another as to the reality of the apparently supernatural; in Cazotte’s text, we simply have conflicting information on what has really happened.

48. Todorov, Fantastic, 171.

49. Todorov’s theory is beset with problems besides his waffling over Cazotte. For example, it is not clear what texts, aside from James’s Turn of the Screw and a few others, really fit into the vaunted category of the “pure” fantastic; this concentration on a more or less empty square means that Todorov passes over without comment the huge variety within “fantastic-uncanny” narratives (from the “supernatural explained” to, say, the “diary of a madman”). And the use of the word “pure” lays bare the extent to which Todorov’s ideology of reading is of a piece with New Critical ambiguity, or with Barthes’s writerly text. By refusing the “irreducible opposition between real and unreal,” the pure fantastic comes to “represen[t] the quintessence of literature” (Fantastic, 167, 168). But quintessences rarely make for accurate historical observation. (For some other points of difficulty, see Brooke-Rose, Rhetoric of the Unreal, 62–71.) Nevertheless, Todorov’s focus on hesitation is important, and I will return to it below.

50. “Diable amoureux, nouvelle espagnole,” 97.

51. Fréron, Année littéraire, 2: 118–19.

52. For example, La Harpe lumps Le Diable amoureux together with the rest of Cazotte’s works, all considered “amusing” enough for one reading—“and in this genre, that is still saying something” (Correspondance littéraire, 1: 389–90).

53. Ampère, Littérature et voyages, 2: 70. Ampère goes on to indict the use of ghosts and the “supernatural explained” as incapable of generating true involvement; he never mentions names, though he must be referencing the English gothic and its French imitators. I will return to the matter of the gothic below.

54. Duvergier de Hauranne, “Contes fantastiques,” 819.

55. Nerval, Oeuvres, 272–73. Nerval’s text was published in article form and used as a preface for a reedition of Le Diable amoureux in 1845; it was later rewritten for Les Illuminés (1852), from whose version I quote.

56. Duvergier de Hauranne, “Contes fantastiques,” 819. “Fiction” here is synonymous with complete poetic liberty to invent.

57. Diderot, Oeuvres esthétiques, 112–13.

58. New, though certainly not without a strong anchor in both its marvelous predecessors and Cazotte’s earlier work. Hunting, “Mille et Une Sources,” has convincingly pointed to the Arabian Nights—mostly the story of Aladdin—as the source for apparitions that cause the protagonist to quake in his boots while he nonetheless holds firm; other stories of the Arabian Nights rehearse the “am I awake or dreaming?” motif. The novelty, then, is the novel’s European setting and the insistence with which its author explores the boundary between this world and the other—an insistence, again, which can easily be taken for properly fantastic (as Hunting does) but is not.

Meanwhile, Cazotte’s earlier work shows that the writer had long been thinking about new possibilities for exploiting a type of naturalized marvelous. His faux-medieval ballad “La Veillée de la bonne femme,” praised by Nerval as the unique French example of gothic, begins: “Spot in the middle of the Ardennes / is a chateau perched on a rock / where ghosts by the hundreds wander” (quoted in Nerval, Oeuvres, 274). Significantly, Cazotte also left early versions of the device that would become a cornerstone of the gothic novel, the “supernatural explained.” Le Lord impromptu (1767) is an “explained” fairy tale: a foundling discovers that the creature he takes to be his good fairy is actually his real mother, who will make his wishes come true by good old fashioned plotting (and lots of disguises). (The short novel, purported to be a translation from the English, bears the oxymoronic subtitle “nouvelle romanesque”: much like Tom Jones [1749], it refreshes the plot elements of romance [romanesque] by bringing them into the present day [nouvelle].) And in his homage to Orlando furioso, Ollivier, Ariosto’s mythical creatures—hypogriffs and such—are reprised as mere mechanical contraptions. Le Diable amoureux is not, then, the only text of Cazotte’s to explore avenues for the marvelous that fairy and oriental tales had not taken.

59. This is in part self-criticism, since an early version of the present chapter, “Permanent Re-Enchantments,” takes just this approach.

60. Todorov, Fantastic, 166.

61. This is the suggestion of Killen, Roman terrifiant, 50; similar parallels are noted by Castex, Conte fantastique, 40–41. A possible link between Cazotte and Lewis has in fact been a subject of debate since the very publication of The Monk; see Peck, “The Monk and Le Diable amoureux,” who demonstrates that influence is unlikely.

62. For the record, Le Diable amoureux was translated into English in 1793 and again in 1798.

63. On the supernatural marvelous of these texts, see note 58 above. As for Otranto in France, the 1767 translation was not reviewed in print, though Grimm reported fairly enthusiastically on it in his manuscript newsletter La Correspondance littéraire. The 1774 reprint was reviewed in L’Année littéraire by Fréron, whom we have seen praising Cazotte; Fréron was negative, and makes fun of the editor’s pronouncement that the first edition had been a success in France. (One additional review in 1774, in the Journal des Beaux-Arts et Sciences, was positive but perfunctory.) The critical fortunes of Otranto, therefore, were not significantly different in France than in England: appreciation was mixed.

64. Walpole, Castle of Otranto, 9–10.

65. The quote is from the second (1778) edition; see Reeve, Old English Baron, 4.

66. Key texts here include Hurd’s Letters on Chivalry and Romance (1762), Montague’s Essay on the Writings and Genius of Shakespeare (1769), and Drake’s “On Gothic Superstition” (1798). “Gothic” in this use is roughly synonymous with “medieval,” and does not designate the modern novelistic use of ghosts and such.

67. Antiquarian interest in the Middle Ages was certainly present in mid-eighteenth-century France; Sainte-Palaye’s 1759 Mémoires sur l’ancienne chevalerie were “the constant source of information about chivalry, not only for Hurd, Percy, and Thomas Warton, but for Gibbon, Joseph Sterling, and later even Byron” (Johnston, Enchanted Ground, 23–24). In contrast to Hurd, however, Sainte-Palaye did not draw from his studies of chivalry a vision of an indigenous, non-neoclassical literature. On French reception of Macpherson, see Van Tieghem, Ossian en France; for another current of French medievalism, see Jacoubet, Comte de Tressan. The “medievalist” vogue in France crested much earlier, with novels such as Lhéritier’s La Tour ténébreuse, d’Auneuil’s Les Chevaliers errants et le génie familier (1709), and Vignacourt’s La Comtesse de Vergi (1722).

68. “Otranto opened the floodgates for a whole torrent of horror-novels. Since Walpole’s time, for a stretch of about 40 years, readers ‘supped full of horrors’” (Varma, Gothic Flame, 42): Varma does not even attempt to back up his assertion by furnishing titles other than the (hardly horrific) Old English Baron and Lee’s (nonsupernatural) The Recess.

69. Hurd, Letters, 120, 102.

70. Beattie, On Fable and Romance (1783), in Clery and Miles, eds., Gothic Documents, 92. Beattie continues: “Mankind awoke as from a dream.”

71. The text of the two reviews can be found in Sabor, ed., Horace Walpole, 68–70.

72. Langhorne’s reviews are reprinted in Sabor, ed., Horace Walpole, 71–72. Clery’s analysis of the articles in the Monthly Review and the Critical Review is perceptive, though she too seems to believe that the reviewers were initially taken in (Rise of Supernatural Fiction, 53–55).

73. Sabor, ed., Horace Walpole, 70.

74. Grimm and Diderot, Correspondance littéraire, 5: 478. Warburton’s mention occurs in a footnote to his Works of Pope (1770), and is reproduced in Sabor, ed., Horace Walpole, 75.

75. “A Gothic story, professedly written in imitation of Otranto, but reduced to reason and probability! It is so probable that any trial for murder at the Old Bailey would make a more interesting story” (Walpole, Correspondance, 27: 381–82).

76. The review, in the Décade philosophique in 1797, is quoted in Killen, Roman terrifiant, 89.

77. Grimm and Diderot, Correspondance littéraire, 5: 478.

78. Paulson, Representations of Revolution; this particular interpretation goes back to Sade, who felt that the works of Radcliffe and Lewis were the result of “the revolutionary shocks which all of Europe has suffered” (Idée sur les romans, 32). Other cultural readings, insisting on more locally English factors, include Sage, Horror Fiction, and Watt, Contesting the Gothic.

79. Brown, Gothic Text. In a somewhat similar vein, see also Mishra, Gothic Sublime.

80. Clery, Rise of Supernatural Fiction, 87.

81. Clery, Rise of Supernatural Fiction, 1.

82. See Moretti, “Slaughterhouse.”

83. Brown, Gothic Text, has located Walpole’s originality in such psychonarration (19–33). For the exploration of thought in the later gothic, see Castle, Female Thermometer, 120–39.

84. My remarks here on hesitation as a device are intended as a hypothesis; serious study of the data on the subject has not even begun. One further hypothesis also deserves consideration: the difference between a device that “works” (hesitation, as first used in the gothic “supernatural explained”) and one that doesn’t (Walpole’s brute supernatural, viewed through the lens of a discovered manuscript; or Cazotte’s variety) may be partially due to chance. In England, Fuller and Smith make use of the supernatural explained; there happened to be a very talented writer, Radcliffe, who saw and exploited the device’s potential, which then led to followers; another talented writer, Lewis, signed on as well, and the rest was literary history. In retrospect, the appearance of this or that fad often seems to be necessary, to make historical sense: we naturally make the gothic novel the expression of a decade of violent revolution, for example. Yet without Radcliffe and Lewis, would there have been a gothic novel at all? And how many other literary fads have fizzled or never occurred—fads about which, had they actually become fads, we would now not hesitate to say they make perfect and necessary historical sense? In this, and like much human endeavor, literary history is very often a profoundly superstitious enterprise: we see meaning where there is only randomness.

85. As reported by Stendhal in his diary and quoted in Trahard, Prosper Mérimée, 104. The uncanny, then, does not burst on the scene so much as it is the result of continual literary changes.

86. Le Diable amoureux was preceded in its first edition by an “Avis de l’éditeur.” The title of this document would lead the reader to suspect the standard pseudofactual feint. In fact, the “Avis,” which mostly is given over to discussing the work’s interesting engravings, explicitly refuses to countenance the “found manuscript” pretense: “[The work] was dreamed in a night and written in a day; it was not at all, as is usual, stolen from the author: the author wrote it for his own pleasure, and a bit for the edification of his fellow citizens” (4). Moreover, there is a good possibility that the “Avis” did actually come from an editor-publisher, or at least from a pen other that that of Cazotte: Fréron, in his review, attributes it to “one [of the author’s] friends, very well known in our capital for his wit and his talent for jest” (Année littéraire, 2: 119).

87. See the remarks of Décote, Itinéraire, 184–86. While unquestionably fluid, nomenclature of the period nonetheless follows patterns, and Cazotte’s does not fit. As for the adjective “espagnole,” it probably derives from the protagonist’s nationality alone, since Cazotte does not appear to be referring to any identifiable Peninsular tradition of supernatural temptation.

88. One might, for instance, argue that since Walpole in his second preface and Reeve in her first acknowledge their own authorship, they use the pseudofactual form fictionally, just like Cazotte. Though I do not think much is to be gained by hair-splitting, the English cases do seem to me a bit more timid: they suppose that we can read unbelievable subject matter only via the cover provided by superstitious times; we are being asked to read over medieval shoulders. But behind this there is a more interesting problem. This book insists that pseudofactual forms are not fictional because they pretend to be true, yet Le Diable amoureux is a good (early) example of a pseudofactual form (the memoir) that admits invention. Isn’t the implication that there are “pseudofactual” pseudofactual novels and “fictional” pseudofactual novels? (If so, Cazotte, possibly along with Walpole and Reeve, would fit into this latter category.) I will come back to this important point in the Conclusion.

CONCLUSION: ON NARRATORS NATURAL AND UNNATURAL

1. Dostoyevsky, Gentle Creature, 60.

2. Cohn, Distinction of Fiction, 37.

3. Smith, On the Margins of Discourse, 30; cited by Cohn, Distinction of Fiction, 20.

4. Searle, Expression and Meaning, 58–75. Genette nuances but more or less follows Searle: “the ‘discourse of fiction’ is in fact a patchwork … of heterogeneous elements borrowed for the most part from reality” (Fiction and Diction, 49).

5. See Hamburger, Logic of Literature; Martínez-Bonati, Fictive Discourse; and Banfield, Unspeakable Sentences.

6. “And as a bitch stands over her tender whelps growling, when she sees a man she does not know, and is eager to fight, so his heart growled within him in his wrath at their evil deeds” (Homer, Odyssey, 2: 275).

7. Cohn, Distinction of Fiction, 10. Rather summarily, Cohn takes this reading of mimesis to be a matter of consensus, though she duly notes the dissenting reading of Gill, “Plato on Falsehood.”

8. Brown, Gothic Text, 41. “Psychonarration” is Cohn’s term for what others call thought reports, inner monologues, and so on.

9. Pavel, “Between History and Fiction,” 23.

10. Epic, for example, was usually thought of as a mixed mode, combining the dramatic (where characters spoke for themselves) and the narrative (in which the poet spoke of them). Because of this mixture, Halliwell remarks, “modern narratological interest in differences of voice, technique, and point of view within third-person narrative has hardly any antecedents in ancient theory or criticism” (Aesthetics of Mimesis, 168n44).

11. See Genette, Narrative Discourse Revisited, 84–95.

12. Culler, “Omniscience,” 24.

13. Culler, “Omniscience,” 32. Given the difficulty of attributing narration in many canonical modern novels to actual persons, some have argued for a narratorless understanding of third-person novels. (These include Hamburger, Logic of Literature, and Banfield, Unspeakable Sentences.) Others, especially Genette (Narrative Discourse Revisited, 96–108), vigorously defend narrative personhood. If we were to rephrase the issue historically, both sides might agree on a compromise: modern literature tempts us to see third-person narration as narratorless in a way that earlier texts do not.

14. Indeed, the twentieth-century novel can and maybe should be understood as the history of “the formal work done [by writers] on the first person” (Lucey, Never Say I, 14).

15. Cohn, Distinction of Fiction, 37.

16. Cohn, Distinction of Fiction, 30–34. Cohn argues that this nonidentity enables the narrator’s limited or biased point of view to become the actual subject of the novel.

17. The challenged posed by the pseudofactual editor to narratological definitions of fiction and nonfiction is also obvious in Genette’s work. In Fiction and Diction (68–79), Genette defines five modes of narrative (they include fiction, autobiography, and history) by altering the identity or nonidentity between three voices—of author, narrator, and character. Editorial pretense, however, wreaks havoc with the schema.

18. Flaubert, Madame Bovary, 81.

19. Perhaps a more interesting counterexample is to be found in the spate of eighteenth-century short novels sometimes known as “it” narratives, that is, tales told by objects such as a coin or a comb. While such narrators are admittedly as unnatural as any invisible stenographer, they are also obvious examples of the venerable rhetorical device of prosopopoeia. “Aristotle says that the narrative poet sometimes transforms himself into something else, not some one else, because the speakers in narrative poems are not always men and women,” writes Renaissance scholar Castelvetro, who goes on to itemize five categories of speakers, of which the last comprises “plants and inanimate objects, like stones, gold, iron, beds, and the like” (Art of Poetry, 30). Prosopopoeia treats animals, objects, or abstractions as natural human speakers, whereas many (though certainly not all) modern narrators are only with difficulty described as speakers at all.

20. Quoted by Palmer, Fictional Minds, 242.

21. Quoted by Palmer, Fictional Minds, 243.

22. Palmer rightly suggests that remarks like Scott’s and Richardson’s prompt the question of how “direct access to fictional minds,” once unnatural, becomes naturalized (Fictional Minds, 243).

23. Pavel, “Between History and Fiction,” 26.

24. “[A]ll literary scholars analyze stylistic structures—free indirect style, the stream of consciousness, melodramatic excess, whatever. But it’s striking how little we actually know about the genesis of these forms. Once they’re there, we know what to do; but how did they get there in the first place? … Concretely: what are the steps? No one really knows” (Moretti, “Novel,” 114)