If Lafayette, therefore, is not a good predictor of the French novel to come, this is mostly because the form she was working with had no future. In studies of the French novel, it has become customary to think of the advent of the historical novella in the 1670s as the key step in the direction of modernity: the extinction of the dinosaur that was historical romance left the terrain open for colonization by the mammalian novel, of which the tiny historical novella was the first example. There is much to be said for this version of events, which is close to the version that Lafayette’s contemporaries told. But in one important matter the hypothesis hides from view the profound commonality between the historical novella and the historical romance: notwithstanding its different structure (it was both short and not built around the exchange of its protagonists’ stories), and notwithstanding its increased emphasis on documentation or its downplaying of physical acts of heroism, the historical novella was perfectly Aristotelian in the type of invention it practiced: it was often (but not always) “more” historical, in that it made greater use of source texts, but at bottom it was historical in the same way. Only pseudofactual forms, which as Lennard Davis has pointed out masqueraded as something like “news,”49 were different: their authors were either the interested parties themselves, or eyewitnesses; they were not poets, picking and choosing between attested versions of illustrious events and then inventing what was needed for the demands of their plot. The pseudofactual was no direct shock to the Aristotelian system, as I’ve noted in the Introduction, since it did not question the preeminence of characters’ reality; but at the same time it represented a distinct mode of composition. From this point of view, then, La Princesse de Clèves tinkers with Aristotelian principles, subverts them even, but in no way opens the door to the regime that will supplant Aristotelian poetics. Subligny’s now-forgotten La Fausse Clélie (1670), the subject of the following chapter, may in fact be much more representative of the novel’s future than Lafayette’s masterpiece.
La Princesse de Clèves looks like fiction: it refuses to assert the existence of its protagonist, who is not even a plausible poetic construction. And it behaves like fiction: it causes readers (or at least Valincour) to envision their relation to that protagonist in a way that does indeed recall later sentimental identification, or better still, the type of engagement proper to the realist novel, which makes accessible to readers the “transparent minds” of its characters.50 But what does it mean to call La Princesse de Clèves fiction? Almost inevitably, such a declaration brings with it a host of erroneous historical corollaries, foremost among them that this novel must therefore be related to later ones, or that it is a first sign of a conceptual change. To say that despite its invented heroine La Princesse de Clèves is not fiction is not to indulge in academic casuistry. It is to remind us that the resemblances between it and later novels are not signs of a pattern but accidents of history. For fiction is not only a sum of characteristics, it is a practice, and practices are communal. From practices individuals can and do deviate, and sometimes in ways that either contemporaries or later generations may consolidate into new practices. But deviations do not always bear such fruit, even if, as with La Princesse de Clèves, they are appreciated, lauded, kept alive in the literary memory. Lafayette’s example does not mean that France knew fiction before Britain did; she did not divine what the novel would look like 150 years later, or recognize, in the midst of general confusion, what fiction “really was.” Her book is a hapax, a one-off, an isolated short-circuit in the Aristotelian machine; Lafayette did not have it in her power to rewire literary creation along the lines of what would become, so much later, “fiction.” If La Princesse de Clèves did not exist, no one would have to invent it; we wouldn’t even miss it, because the history of the French novel would look more, not less, comprehensible in its absence.
Appendix: History and the Historical Novella
To Valincour’s objection regarding Lafayette’s “visibly false” heroine, Charnes responded with the argument that La Princesse de Clèves belonged to a “third species” of modern novels of whose compositional principles Valincour was ignorant. The move was astute, rhetorically speaking, since it allowed Charnes to cast his adversary as a pedant behind the times. As I’ve remarked, however, Charnes’s taxonomy was pure sophistry: there was no contemporary species of novels that, like La Princesse de Clèves, advertised the nonexistence of protagonists. This appendix provides justification for my claim that in spite of the fact that historical novellas of the period displayed much variation in their use of history, Lafayette’s contribution remained sui generis.
Knowing precisely what to count as a nouvelle historique is not necessarily an easy matter, even if the generic tag was applied by contemporaries—appearing almost simultaneously with the genre itself, on the title page of Saint-Réal’s Dom Carlos (1672). Roughly twenty works from the last three decades of the seventeenth century would identify themselves thus. But there are complications. First, works with this subtitle do not always designate the same thing. Generally, the historical novella was set in the past; its protagonists were historical in the sense that they were attested in histories, memoirs, genealogies, and so on. On occasion, however, the tag was applied to works set in the present. Second, many works that bear the hallmarks of the genre (including both Lafayette’s efforts) have no such subtitle; in rare cases, an alternate designation such as histoire tragique takes its place. Third, although a title containing a proper name and aristocratic title often announced a historical novella, in some cases, especially in the genre’s infancy, novels bore more general titles—as in Villedieu’s Annales galantes (1670) or Désordres de l’amour (1675); and conversely, the titular proper name could announce “true stories” set in the very recent present—as with Boursault’s Le Marquis de Chavigny (1670), Salvan de Saliès’s La Comtesse d’Isembourg (1678), or Du Plaisir’s La Duchesse d’Estramène (1682). Titles alone, then, are only a rough guide to the actual content of these novels. Searching this content for works set in the past leads to a list of a bit more than fifty historical novellas published in the last three decades of the century, at which time the genre more or less went into eclipse.51
How historical, precisely, was the historical novella? To readers of Lafayette, or those familiar with the extremely careful use of sources evident in the other commonly read examples of the genre, Dom Carlos and Les Désordres de l’amour, the answer would probably be very: nothing could seem farther from the vague historicity of d’Urfé’s Gaul or Scudéry’s Rome than the patient marquetry of Lafayette, Saint-Réal, and Villedieu, in which sources of different types—memoirs, genealogies, official historiography, even manuscript material—are fitted together. Yet the divide between the two may not have seemed quite so unbridgeable to contemporaries. Of about twenty surviving works of the period that incorporate the generic designation into their title, and even if we exclude outlying quirks, we find texts that vary widely in terms of the period treated and the amount of historical detail provided.52 The production of 1678 can serve as representative; this was the year in which the most historical novellas were published, including of course La Princesse de Clèves. Six of nine historical narratives that year were subtitled nouvelle historique. Two took place in the early Middle Ages, and far from France: Le Comte Roger, souverain de la Calabre ultérieure, whose author is known only under the initials “L.L.B.”; and Alfrede, reine d’Angleterre, attributed to Antoine Torche. If the main lines of the biography of Roger Guiscard (c. 1016–85) were fairly well known, the number of sources was very limited, and