It almost goes without saying that a new book by Michael Riffaterre is an important book. Riffaterre's distinguished reputation virtually guarantees that all his writings will command immediate attention and respect. And Fictional Truth does not disappoint: although brief in compass it is large in conception and vast in its implications. Its language is succinct and its vocabulary precise, although not necessarily easy to digest. Its very conciseness and precision make great demands on the reader. Even those with considerable experience with narrative theory will have to study it carefully to begin to make sense of it. It is as dense as any scientific treatise and requires patient and repeated reading.
One difficulty almost every reader will face is that Riffaterre uses the term "truth" in a deliberately provocative way. Although not excluding or abandoning completely the notion of referentiality, he reconceives it along almost Derridean [End Page 665] lines so as to bring it within the domain of pure semiosis. Signs indeed make reference, he argues, but not to anything outside the structure of signs: "In fact, exterior referentiality is but an illusion, for signs or sign systems refer to other sign systems: verbal representations in the text refer to verbal givens borrowed from the sociolect, but such verbal givens are actually present in the text, explicitly or implicitly, as presuppositions." In other words, il n'y a pas de hors-texte. The recognition of truth, then, is not a matter of matching words to experience because "neither experience nor previous readings are needed, only linguistic competence: truth is nothing but a linguistic perception." These will doubtless be fighting words in some quarters.
But not for me. I find myself both persuaded and at times dazzled by Riffaterre's keen insights, and I realize that, although I was never his student, I remain in many ways his pupil. He produces theory, but it is a practical theory that always arises out of close attention to particular texts. It is the kind of work I wish I could do as well, and I find myself surprised that I am troubled by even one (perhaps minor) point. Although I happen to agree in principle with Riffaterre's semiotic definition of truth, I am uncomfortable with the inability of his vocabulary to distinguish among various forms of intra-linguistic (or intra-semiotic) validation, all of which he calls "truth." A number of the techniques Riffaterre says realistic writers such as Proust use to establish such "truth" are in fact hardly distinguishable from ways in which aggressively unrealistic writers win assent for outlandish tall tales and other "lethetic fictions" (my terminology in The Incredulous Reader) that have no interest whatsoever in establishing themselves as "true." There is a difference between winning rhetorical assent and establishing truth, and Riffaterre's terminology obliterates this distinction. This is not to deny that "truth" is a form of rhetorical assent: it is merely to point out that it is not the only one.
Under other circumstances, Riffaterre might have called his book The Rhetoric of Fiction. It would have been perhaps the most fitting title of all, for what is at issue is indeed the way in which fiction marshals the resources of language to make itself convincing. But Riffaterre uncovers aspects of fictional rhetoric that Booth did not address, delving into the micro-structure of even such an expansive novel as À la Recherche du temps perdu to show how it holds our interest and wins our assent from word to word and sentence to sentence. Fictional Truth is essential reading for everyone interested in the way narrative works. [End Page 666]