On the Very Idea of a Definition of Narrative: A Reply to Marie-Laure Ryan

I

"Quid est ergo tempus?", St. Augustine famously asked. "Si nemo ex me quaerat scio; si quaerenti explicare velim, nescio". These words, which strike me as some of the most deeply honest and self-aware of his Confessions, describe an experience that must surely be familiar to most narratologists: Augustine's philosophical musings on the nature of time lead him to conclude that when no-one is asking him what time is, he knows, but as soon as he attempts to explain it, it turns out that he doesn't know. So it is with narrative. Mutatis mutandis, we all know what a narrative is: we all recognise one when we see one. But when we try to commit our knowledge to paper, it inevitably turns out that for every generalisation there is an exception, for every taxonomy there is a misfit, and for every definition there is always room for further definition, as extraneous elements creep into our classifications. Such, no doubt, is life.

The question then arises as to the consequences of a knowledge that is simultaneously present and absent. It would seem odd (or, better, wrong) to claim that we don't know what a narrative is if our definitions of it are unsatisfactory. Is anyone going to claim that Augustine didn't know what he was talking about in his meditations on time? And yet Augustine freely confesses that he both does and does not know what time is. Like so many of us—and Marie-Laure Ryan's response to my work has convinced me that I am indeed just such a person—Augustine wants to have his cake and eat it. But is this a problem?

I see no compelling grounds that suggest it is, and I am happy to concur with Ryan, herself concurring with Bal, that "Usually, we don't doubt, don't wonder, about the status of a text, and watching a film we may or may not be carried away, but hardly does it seem important to ask if what we are seeing is in fact a narrative" (194). Wrangling about the kind of terms we use in defining narrative is simply not an activity that most of us bother with: we take our knowledge of such things for granted, and, to pick up my earlier example that Ryan objects to, let me reassert that most readers of narrative can differentiate between a cartoon and an instructions manual as easily as they can differentiate between a hawk and a handsaw.

I am indebted to Ryan's flatteringly thorough reading of my piece for pointing out to me an ambiguity in my position—she calls it a contradiction—and all the more grateful to her for affording me this opportunity to set the record straight. My paper "From Narrative Representation to Narrative Use" sought to demonstrate that the terms with which most leading narratologists have defined their subject matter do not in and of themselves provide a basis for differentiating between what is and isn't narrative, because they are based on features—representation, sequentiality—that inhere in texts we would justifiably consider non-narrative. Ryan, however, points out that my paper attacks and undermines the traditional distinction between narrative and non-narrative while itself making use of the designations "narrative" and "non-narrative". This does indeed seem like a contradiction, which I hope to unravel here.

The contradiction Ryan diagnoses is characterised as a contradiction between a traditional view of narrative as a constant, and a different view of narrative as a variable, based on the uses to which the text is put. As Ryan says, "Narrative theory can have it one way or the other, but making narrative into both a variable and a constant is plain contradiction" (192). This is an astute observation, and would indeed be a fatal flaw in the paper had it actually argued that narrative is a constant. It makes no such assertion. It does assert that there are various genres and forms of discourse that tend to be designated as "narrative", but...