8: The Figure of the Unreliable Narrator
tations based on the incoming information about the apparent trustworthiness of their sources. The interplay between the unpredictability on the one hand, and the unavoidable regularities of our information-processing cognitive systems, on the other, is what makes it possible for such writers as Cervantes, Lennox, Dostoyevski, and Nabokov to play with us in millions of ever new ways, degrees, and combinations, and it is what ensures that the game of fiction is still going strong after thousands of years.
SOURCE-MONITORING, ToM, AND THE FIGURE OF THE
UNRELIABLE NARRATOR
I am pleased to report that there seems to be an affinity between my
JL take on an unreliable narrator as a function of textual experimentation with our source-monitoring and the view first introduced by Jonathan Culler and then explored by Monika Fludernik and Ansgar Nunning. As Nunning has argued recently, the critic may account
for whatever incongruousness s/he may have detected by reading the text as an instance of dramatic irony and by projecting an unreliable narrator as an integrative hermeneutic device. Culler . . . has clarified what is involved here: "At the moment when we propose that a text means something other than what it appears to say we introduce, as hermeneutic devices which are supposed to lead us to the truth of the text, models which are based on our expectations about the text and the world."1 . . . [Similarly, as] Fludernik has shown so convincingly [in a related context,] the projection of an unreliable narrator can be seen as "a result or effect of the reader's pragmatic interpretation of textual elements within their specific literary context."2
At the same time—although I agree with Nunning that we import the figure of an unreliable narrator because we need to frame in familiar social terms a perceived pattern of textual ambiguities—I am less troubled than he is by all the anthropomorphizing that goes into it. From the point of view of cognitive theory, it is not terribly surprising that we conjure up an extra mental presence when we intuit that the narrative is monkeying around with our source-monitoring capacities. There is a very short step— thanks to our Theory of Mind, which is ever hungry for more material to work on—from starting to suspect that the text is fooling us, to ascribing a whole host of other states of mind to that wily typographic entity. As Uri Margolin observes in a different context, " [S] ince we cannot but conceive of narrative agents as human or human-like, it is a basic cognitive requirement of ours that we attribute to them information-processing activities and internal knowledge representations."3
It is not accidental, then, that Phelan's recent exploration of unreliable narrators, Living to Tell about It, describes different types of fictional unreliability in terms of specific behavioral patterns on the part of the text as well as on the part of its readers. Phelan defines unreliable narration thus:
Narration in which the narrator's reporting, reading (or interpreting), and/or regarding (or evaluating) are not in accord with the implied author's. There are six main types of unreliable narration: misreporting, misreading, and misregarding, underreporting, underreading, and under-regarding. The two main groups can be differentiated by the activity they require on the part of the authorial audience: with the first group— misreporting, misreading, and misregarding—the audience must reject the narrator's words and reconstruct an alternative; with the second group—underreporting, underreading, and underregarding—the audience must supplement the narrator's view.4
Indeed, given what we know about the workings of our metarepresentationality and Theory of Mind, we should be able to have our theoretical cake and eat it too. That is, I have no qualms about thinking of the unreliable narrator as first and foremost a function of the textual engagement of our cognitive adaptations for source-monitoring while at the same time appropriating, for the rest of my argument, Phelan's excellent (and anthropomorphic) classification.
A given narrator, Phelan observes, "can be unreliable in different ways at different points in his or her narration."5 For example, Frankie, the child-narrator of Frank McCourt's Angela's Ashes, both misreads and misregards what is going on around him, whereas Nabokov's Humbert misregards, misreports, and underreports his actions and Lolita's responses to them. Likewise, Mr. Stevens, the protagonist of Kazuo Ishiguro's Remains of the Day, misreports and misregards certain events of his story as well as underreports and underreads his own motives.6 When "Stevens says that 'any objective observer' will find the English landscape 'the most deeply
9: The Implied Author
satisfying in the world,' he demonstrates a misperception analogous to his saying that 'any objective observer' would find the English cuisine the most satisfying in the world."7 He thus exhibits unreliability both "on the axis of knowledge/perception" (misreading) and "on the axis of ethics and evaluation" (misregarding).8
Note that Stevens is perfectly sincere in his belief about the "objective" superiority of the English landscape. This belief (Phelan suggests) might be rooted in his "mistaken value system,"9 which takes as a given certain subjective assumptions about the world. From the point of view of cognitive theory, by considering his take of the English landscape universal, Stevens loses track of himself as a source of this representation. To "reject the narrator's words and reconstruct an alternative,"10 the reader thus has to become aware of the missing source tag—"Stevens thinks that . . ."— and to reapply it.
Every step in this process engages and titillates our metarepresentational capacity: We come into the awareness of the missing source tag. We reapply the tag. We contemplate various ramifications of the difference between the two representations ("the English landscape is the most deeply satisfying in the world" vs. "Stevens thinks that the English landscape is the most deeply satisfying in the world") that jostle against each other in our readerly consciousness. We begin to wonder what other representations within the story may also be missing their source tags. As Phelan points out, "[OJnce any unreliability is detected all the narration is suspect"11—in some narratives, the game of the missing source tags is never really over. We close such books with a strange feeling that the state of cognitive uncertainty that they induced in us may never be fully resolved. We continue guessing which representations within the story deserve to be treated as "true" and which have to remain metarepresentations with a source tag pointing to the first-person narrator.
9
SOURCE-MONITORING AND THE IMPLIED AUTHOR
similar kind of guessing game (whom to trust and how much?) takes
place when literary critics contrast such figures as the "real" author of the text, its narrator (especially when unreliable), and its "implied author." If you are unfamiliar with the latter term, Gerald Prince's Dictionary of Narratology defines it as "the implicit image of an author in the text, taken to be standing behind the scenes and to be responsible for its design and for the values and cultural norms it adheres to."1 Students of narrative have been debating the added value of this concept at least since the publication of Wayne Booth's The Rhetoric of Fiction, in which Booth suggested that the category of implied author captures our "intuitive apprehension of a completed artistic whole" of a given text.2 (For a good summary of critical grapplings with "this anthropomorphized phantom," see Nunning.3) Among cognitive narratologists, Palmer remains skeptical of the possibility of maintaining the distinction between the narrator and the implied author, observing that though "clear in the case of first person narrators, [this distinction] can be problematical in other cases." As Palmer sees it, when it comes to practical discussion of many novels, it is not even possible to maintain a "coherent distinction between the agency that is responsible for selecting and organizing the events (as Prince describes the role of the implied author), and the voice that recounts them (the narrator)."4
On the one hand, I see Palmer's view as broadly corroborated by what we are learning about our metarepresentational capacity. Constantly keeping track of the difference between the implied author and the narrator means in effe
ct retaining a source tag behind every minute instance of narration and, moreover, doing so after you have already bracketed the whole story as a metarepresentation pointing to the author. It means, for example, saying to yourself as you read Pride and Prejudice and come across Lydia Bennet's elopement with Wickham: "Austen claims that Lydia ran away with Wickham"—a kind of micro source-tracking5 that is simply too cognitively expensive and as such is not a default mode of our reading process. It seems to me that it is precisely because we do not, in our everyday reading practices, trace back to the author every single representation contained in the text (once we have bracketed off the whole fictional text as a metarepresentation) that the writers fond of unreliable narrators can play their complicated games with their readers.
On the other hand, I am not really that invested in debating either the overall usefulness of the category of the implied author or the cognitive feasibility of maintaining such a category. What I find more fascinating is the cultural history of the figure of the implied author. Susan Lanser has characterized its introduction into the narratological discourse in the early 1960s as a "problematic compromise." As she sees it, the figure of the implied author "not only adds another narrating subject to the heap but it
9: The Implied Author
fails to resolve what it sets out to bridge: the author-narrator relationship." At the same time, at a "crucial moment in literary history the word 'implied' did provide a respectable prefix with which the mention of the author became permissible."6
Lanser is referring to the time when the terms author and reader had "all but disappeared from the analysis of point of view, because they were not considered properly textual personae." By 1960, "Anglo-American New Criticism had taken as a basic tenet the autonomy of the text as a concrete linguistic object; thus it became virtually taboo to speak of the text as an act of communication among real people in the real world."71 find Lanser's account of the compensatory function of the term implied author particularly gratifying from the cognitive perspective I champion here. It appears that, prevented from speaking about the "real" author behind the fictional narrative, critics nevertheless found a way of still retaining a source tag behind this narrative—by introducing the category of the "implied" author. Compare this act of cognitive compensation to the one that I described earlier, when talking about Barthes's-Foucault's concept of the "Death of the Author." Here, the author is substituted by the "implied" author; there, the author was substituted by the reader. It seems that a culture will find a way to insinuate a source tag into its perception of a representation that is a metarepresentation. A work of fiction has to have an agent-specifying source tag affixed to it, however extravagant (e.g., "dead" or "implied") that agent may seem at certain historical junctures.
Let us give this screw yet another turn. The implied author is alive and well and entering his/her forties. As a heated discussion at a recent meeting of the Society for the Study of Narrative Literature demonstrated, this figure continues to exercise a strong pull on critics' imagination, with some of my colleagues questioning our need for such a concept and others reaffirming its usefulness. I am tempted to see such debates as a function of the source-monitoring ability played out in a very particular social environment, that is, among the people self-selected to pay attention to textual ambiguities. Here is how it works:
Our source-monitoring adaptations are generally on the lookout for material to work on, ready to seize on any evidence that a given representation could be processed as a metarepresentation. Nothing is sacred, nothing is safe from being turned from "truth" to a representation accompanied by a source tag and thus processed under advisement ("Chocolate is good for you." "Says who?"), although different ideological climates may actively encourage some types of metarepresentational processing and discourage others. Combine this general cognitive tendency with the professional training of a literary critic, and it is not unlikely that this individual would be more attuned to the possibility of seeing not just one source behind Pride and Prejudice (i.e., Jane Austen) but a rich hierarchy of sources (i.e., the "real" Jane Austen, the "implied" Jane Austen, the "narrator" of Pride of Prejudice, and so forth). In other words, whereas our shared cognitive adaptation for source-monitoring makes it in principle possible both for me and for my first-year students to see those multiplying authors behind the text, it might be easier for me than for them (at least initially) to achieve such a split vision. I think of the relative ease with which it comes to me as my cognitive-professional hazard.
-10
RICHARDSON'S CLARISSA: THE PROGRESS OF THE
ELATED BRIDEGROOM
hen I think of fiction and cognition in literary-historical terms,
attempting to reconstruct, in particular, the development of the motif of the "Quixotic" imagination from Cervantes to Nabokov, I inevitably return to Samuel Richardson's novel Clarissa (1747-48). Clarissa has been deservedly admired by numerous literary critics, and it is currently going through a pedagogical renaissance, being increasingly taught, even in its forbidding 1,500-page entirety, in a variety of graduate and undergraduate college courses. With the advent of a "cognitive" approach to literature, however, it also ought to be acknowledged as a massive and unprecedented-in-Western-literary-history experimentation with the readers' Theory of Mind and metarepresentational ability, experimentation that certainly made possible the later-day mind-games played by Lolita and Pale Fire}
In this section I argue that in Clarissa, Richardson created a kind of protagonist that we today would call an unreliable narrator. I follow a series of episodes in the novel that increasingly force the reader to doubt the trustworthiness of at least one of its two narrators, and I discuss the cognitive effects of being confronted with a character who seems to believe his own lies. I suggest, in particular, that the presence of such a personage induces in us a state of metarepresentational uncertainty, thus providing a rich stimulation for our Theory of Mind.
10: Richardon's Clarissa
FIGURE 2. Clarissa dying. Reproduced courtesy of McMaster University Library.
(a) Mind-Games in Clarissa
Clarissa is a story of two brilliant young people, Clarissa Harlowe and Robert Lovelace, fatally misreading each other's minds in the course of their deeply troubled courtship. Lovelace, a paradigmatic eighteenth-century "rake," committed to seducing and subsequently abandoning incautious virgins, and Clarissa, a paragon of beauty, piety, and foresight, live out the eighteenth-century version of the "ultimate challenge": will Clarissa convert Lovelace to her exalted system of values and prove that "the reformed rake makes the best husband," or will Lovelace sweet-talk, cheat, and intimidate Clarissa into cohabitation without marriage, his "darling scheme" and his sign of "triumph" over the whole female sex and their pretensions to "virtue"?
Structured as a series of epistolary exchanges between Clarissa and her confidante, Anna Howe, Lovelace and his confidante, John Belford, and the occasional letters from and to their respective families, the novel is simultaneously claustrophobic and boundless. The protagonists are mostly confined to their writing-desks, reporting to their respective friends in painstaking detail their endeavors to guess, second-guess, plant, anticipate, and interpret each other's thoughts. The outcome of this obsessive mind-reading is such that Clarissa and Lovelace stop communicating altogether and die, or, rather, commit what could be considered thinly veiled acts of suicide. Clarissa wills herself to die (figure 2), possibly out of commitment to her developing view of herself as a tragic heroine—indeed, a martyr—who moves inexorably toward her terrible and instructive end,2 possibly from depression induced by Lovelace's manipulation of her reality, a manipulation that makes her feel that none of the mainstays of her moral world—familial love, compassion of the strong for the weak, communal ties—can survive when confronted with playful but determined evil. Lovelace dies because he came to be so emotionally invested in her that he cannot go on after she passes away.
Before I turn to examining Richard
son's experimentation with our metarepresentational capacity, let me make a point that may sound like old news to you by now. This point, however, cannot be repeated often enough in a book that hopes to put the cognitive-evolutionary concept of the Theory of Mind on the map of contemporary literary studies. Clarissa and Lovelace may be preternaturally adept at planning and deflecting each other's mental gambits, an intellectual one-upmanship that marks them as exceptional among other characters in the novel and justifies the appellations of "genius" generously bestowed upon them throughout the narrative. And yet the truly amazing and sustained feat of mind-reading takes place not when Clarissa "sees through" Lovelace's new contrivance or when Lovelace anticipates her seeing through it and prepares a plan B. It takes place when we as readers of Richardson's novel attribute the generous capacity for thoughts and desires to each fictional character, however tenuously delineated, and then proceed to interpret his or her behavior in terms of his or her underlying mental world, supplying a myriad of absent
10: Richardon's Clarissa
links, assumptions, and tacit explanations that allow us to see the story as a rich and emotionally coherent whole. In our interactions with Clarissa (or any other work of fiction), we take our own mind-reading capacity completely for granted and notice it no more than we notice oxygen when we wake up in the morning, an obliviousness which does not, however, render either oxygen or our ToM less important for our everyday life.
Back to Clarissa and metarepresentationality. One of the central premises of Richardson's novel is that its male protagonist is a consummate liar. The intellectual one-upmanship between Lovelace and Clarissa that I have just mentioned is set in motion by his constant endeavors to deceive her. He plots behind her back to set her own family against her; he introduces to her as seemingly respectable people a bevy of prostitutes and criminals; he forges her letters; he dons disguises and draws unsuspecting strangers into assisting him in tricking her.
WhyWeReadFiction Page 11