Some Problems with the Concept of the Narrator in Bortolussi and Dixon's Psychonarratology

An important recent contribution to the study of narrative and, specifically, literary narrative, is Marisa Bortolussi and Peter Dixon's Psychonarratology: Foundations for the Empirical Study of Literary Response. The authors combine insight and intuitions from narratology ("classical" and later developments) with research methodology from discourse and cognitive processing of narrative in order to give an "empirical turn," and push, to the study of narrative. Bortolussi and Dixon's main contention is that literary response involves real readers' constructions, as opposed to "ideal" or "virtual" readers (i.e., undifferentiated, globalized constructs of readers) in narratology, and that real readers create their constructions from objective features in the text. The main emphasis throughout the whole book is on the distinction between features, "anything in the text that can be objectively identified," and constructions, "events and representations in the minds of readers" (28). The authors suggest that features may be identified in texts according to several criteria as being "objective, precise, stable, relevant, and tractable" (38), and they explain what each criterion means. As such, features are manipulable for purposes of research on real readers and also, by implication, there are constraints on readers' constructions from them—in other words, "there is a text" in Bortolussi and Dixon's class and "not everything goes."

Bortolussi and Dixon's basic assumption is that a text precedes and is responsible for readers' reactions. Readers' constructions, however, cannot affect textual features (52). In other words, there is a unidirectional interaction between readers and texts, from features to readers' constructions.¹ As for reader constructions, the aspect Bortolussi and Dixon emphasize is the processing of the text: the cognitive and mental representations for the purposes of coherently understanding such texts. Bortolussi and Dixon's discussion is systematic and focuses on five major narratological concepts, but for the purposes of my discussion I will concentrate on the narrator, the central concept they consider as interacting with and influencing all other significant concepts they discuss. In emphasizing this concept they continue the long tradition of discussions of narrative that insist on the mediating role, especially in fiction, of the narrator. In line with their approach, it is clear that "readers must create a representation of the narrator" (60), but of even more significance for them is that "the reader may represent the narrator as if the reader and the narrator were participating in a communicative situation. The presence, in the mind of the reader, of this communicative situation colors virtually all aspects of the text and its interpretation" (60). Though in Psychonarratology they incorporate previous analytical concepts of the narrator, according to Bortolussi and Dixon the earlier narratological analyses do not contain "explanations regarding how textual clues interact with reader's goals, experience, expectations, and so on in constructing representations of the narrator" (66).

Their main disagreement with "classical" narratological analyses has to do with the communicative model on which it is based, and which they briefly survey (67-68), and the distinctions between real and implied author, and between implied author and narrator. Their point of contention is that the reader referred to is always a hypothetical, undifferentiated construct "extrapolated from the textual features" (69) and not a real one. As such, this reader will always respond like a narratologist, one of whose interests is a taxonomic, descriptive identification of types of narrators. Bortolussi and Dixon specifically mention their rejection of the "conventional narrative communication" (69) model. This rejection is based on the fact that "[t]he reader and author do not share common perceptual ground, they cannot engage in communicative processes of confirmation and error correction, and the same text can be processed quite differently depending on the context, the knowledge and goals of the reader" (74). Instead, they propose a narrator who is "a mental representation in the mind of the reader . . . similar in many respects to that which one would construct of a conversational participant" (72...