In Defense of Definition
This text is the fourth installment in an exchange that began with an article by David Rudrum, “From Narrative Representation to Narrative Use: Towards the Limits of Definition,” published in Narrative 13.2, May 2005.
I commented on the article in “Pragmatics, Semantics, and Narrativity: A Response to David Rudrum,” Narrative 14.2, May 2006, and Rudrum responded, in the same issue, with “On the Very Idea of a Definition of Narrative: A Reply to Marie-Laure Ryan.” Since the protocols of the journal give the last word to the original author in a dialogue thread, I offer my response to Rudrum’s reply on my private Web site.
For the sake of brevity I will not turn these comments on Rudrum’s response into a litany of “Rudrum misreads me when he says that I say,” and I will concentrate instead on the positive claims he makes in his reply. Rudrum and I agree on at least one point, namely that there are too many different uses of narrative to incorporate them into a definition. I conclude that definitions should be grounded in semantics rather than pragmatics, more precisely in the mental representation that users form when reading or watching a narrative. In his response Rudrum draws a different conclusion, one that is not at all evident in the original article: because the uses of narrative are too varied to take them all into account, we shouldn’t try to define narrative. While the subtitle of the original article, “Towards the Limits of Definition,” does indeed express skepticism with regard to the possibility of definition, the last sentence—the one that is supposed to summarize the thesis of the paper, and that lingers the longest in the reader’s memory-- is clearly suggesting that Rudrum is trying to improve existing definitions by locating narrativity in social conventions: “As long as narratology remains tied to [a conception of narrative as representation], and tied to a philosophy of language that foregrounds signification above and before questions of use and practice, it seems that a satisfactory way of defining and classifying its subject matter will continue to elude it” (202-203). If there is a way to have one’s cake and it too, as Rudrum admits he does, it is indeed to argue both ways, and I think that I can be excused for taking up one of those leads rather than the other. Let me now respond to the new (or newly clarified) claim.
In the response Rudrum reiterates that to narrate is to engage in a language game, but he now invokes Wittgenstein’s claim that the concept of game is nothing more than a family-resemblance notion as a justification for not attempting a rigorous definition of the idea of narrative game. Not only does Rudrum see definition as impossible, he regards it as dangerous: “[Ryan’s proposal] carries the danger inherent to all definitions--that of setting a view of the subject that is at best narrow, and at worst foregrounds certain kinds of texts over and above others.” Or: “[definition] carries with it the risk of valorising certain texts above others in the narratological hierarchy—a risk, I contend, that is is inherent to any kind of definition.” These remarks are a little bit surprising from somebody who rejects classical definitions because they fail to exclude graphic model-airplane assembly instructions from the domain of narratology. Obviously there are some texts, or rather, some uses of text that Rudrum wants to foreground; what he does not want to do is commit himself to offering criteria that support his decisions. Without exclusion and foregrounding, narratology becomes a “theory of everything,” which really means a theory of nothing.
Rudrum denies that there is any feature that appears in all uses of narrative (hence his dismissal of my use of the word constant), just as Wittgenstein denies that there is any single rule that appears in all games. But game theorists such as Johan Huizinga, Roger Caillois and Bernard Suits will tell you that the common features of games do not reside in their particular rules, but in the fact that they are all rule-governed activities in which we engage for the sake of pleasure. (The further specifics of their definitions do not matter here; what matters, is that they don’t accept Wittgenstein’s pronouncement as the last word on the subject.) As for narratologists, they may tell Rudrum that what is common to all uses of narrative is just that: the use of narrative, which leads us back to square one, for how do we know that a behavior is a use of narrative if we cannot say anything about the nature of narrative ? Without a common denominator, the signifying practices that form the concern of narratology become a heterogeneous collection reminiscent of a humorous taxonomy attributed by Jorge Luis Borges to a Chinese encyclopedia. This taxonomy divides animals into: “(a) belonging to the Emperor, (b) embalmed, (c) tamed, (d) sucking pigs, (e) sirens, (f) fabulous, (g) stray dogs, (h) included in the present classification (i) frenzied, (j) innumerable, (k) drawn with a very fine camelhair brush, (l) et cetera, (m) having just broken the water pitcher, (n) that from a long way look like flies.” 
Rudrum does not want narratologists to do what intelligence tests routinely ask for, and what arguably constitutes the purpose of science: find the regularities and commonalities that lie behind the diversity of the phenomena under investigation. In defense of his rejection of definition, he observes that “the vast majority of contemporary literary criticism carries on without overmuch regard to these definitions.” While it is probably true that a literary critic can produce an insightful reading of a story without giving much thought to the nature of narrative, I would venture to say that a literary critic is neither a narratologist nor a literary theorist. As the need to create special narratology sessions at the conferences of the Society for the Study of Narrative Literature has made clear, not all readings of narrative are narratological. The difference between literary criticism and narratology lies in the fact that the former is concerned with the particular while the latter, as a theory, is concerned with the general: and what is a definition, if not an attempt to reach the highest level of generality ? To borrow Rudrum’s examples, I don’t see how geologists could conduct meaningful research without at least a working definition of rocks (without which they could mistake tree trunks for minerals), nor how anthropologists could identify a primate skeleton as a human ancestor without a precise idea of human physiology. What Rudrum seems to overlook, in the last analysis, is the heuristic value of imperfect definitions. He has traded the risk of proposing inevitably controversial definitions for the risk of denying any value to taking the risk inherent to definition.
 Quoted by Michel Foucault in the preface of The Order of Things: An Archeology of the Human Sciences. London: Routledge 2002 (xvi). Foucault does not identify the source.
Return to Marie-Laure's Home Page