1 : INTRODUCTION
ters and events do not primarily reproduce people and situations in life. Thus, it does not make sense to say that the depiction of a particular fairy or talking pig is realistic or unrealistic. Antimimetic authors can also parody these forms as well, as we see in Angela Carter's postmodern rewrites of classical fairy tales. In my sections I will be concerned with narratives that are predominantly and, in fact, flagrantly antimimetic, since I find antimimetic texts more challenging than nonmimietic narratives in the ways they contest the conventions of nonfictional and realistic representation. [BR]
20. For further reading on these issues, see Alber and Heinze; Alber, Iversen, Nielsen, and Richardson (2010); Richardson, "Narrative Poetics"; and the Unnatural Narratology website homepage http://nordisk.au.dk/forskning/forskningscentre/nrl/ unnatural/. [BR]
I
2
Authors, Narrators, Narration
James Phelan and Peter j. Rabinowitz
I
"Somebody tells." This chapter will begin to sketch out the rhetorical tangles around the apparently simple phrase that begins our definition. Reading literary narrative is typically more difficult than reading, say, physics textbooks because the "somebody" who tells is neither simple nor stable—and because the resulting complexity and instability are not weaknesses to be minimized but key components of the experience, including the particular pleasures it brings.
Most obviously, from the rhetorical perspective, in many texts (including all fictional narrative) the somebody who tells is split into at least two figures, the author and the narrator. There's little need to rehearse this distinction here: its something pointed out in nearly every manual on narrative fiction, and it's something that children learn fairly quickly. Only the most unschooled reader doesn't understand that Huck is not Twain. But while the distinction itself is relatively unproblematic, there are complications with each of the terms. Furthermore, once we shift our attention to "narration," we need to recognize that authors often use the dialogue between characters to accomplish telling functions.
AUTHORS
Surprisingly, it's the concept of "author" that nowadays generates the greater
2: AUTHORS, NARRATORS, NARRATION
JAMES P H E L A N AND PETER J. RABINOWITZ
theoretical wrangling. Two issues are particularly vexed. First, especially in the wake of poststructuralisms provocative proclamation of "The Death of the Author," there is serious disagreement about whether we should be talking about authors at all, especially about authors as . . . well, authorities on or even designers of texts whose designs are of any significant consequence for interpretation—if they can be recovered at all. As committed pluralists, we believe that the question ought not to be posed as an either/or: valuable critical work can be done by ignoring authors and their purposes, and just as valuable work can be done by attending to them. A purely formal description such as Vladimir Propp's account of the grammar of Russian folktales, can be highly illuminating without any discussion of authorial participation. So too can an ideologically oriented analysis such as D. A. Millers Foucauld-ian account in The Novel and the Police of Victorian fictions role in enforcing certain cultural prescriptions. But to the extent that you are considering narrative as a communicative process, then authors, and their communicative purposes, matter: there can be no rhetoric without a rhetor. Our commitment to pluralism interacts with our commitment to rhetorical theory in our claim that viewing narrative as a communicative process is one valuable (and widely practiced) way of thinking and that certain consequences follow from looking at narrative in this way.
In stressing the authors decisive role, however, we are not suggesting that the task of interpretation (or the goal of reading) should be reduced to the discovery of the author's conscious intentions. As we noted in the introduction, we account for the effects of narrative by reference to a feedback loop among authorial agency, textual phenomena, and reader response. Furthermore, the exercise of authorial agency entails necessary sacrifices: as Wayne C. Booth—whose 1961 book The Rhetoric of Fiction persuasively demonstrated the advantages of approaching narrative as rhetorical action— eloquently argues, the choice to do one thing inevitably blocks your ability to do other things. For example, once Twain decides to tie Huck's growing ethical maturity to his ability to negotiate the conflict between his feelings for Jim as a man and his sense of how whites act toward slaves, Twain has to set his novel before the Civil War, which in turn means that there are certain of his favorite themes (for instance, his criticisms of postwar economic excesses) that he cannot take up. Then, too, there can be unintended (or sometimes unanticipated) consequences of some things an author chooses to do: as Ralph Rader has pointed out, certain "indirect facts" of the text, certain aspects of a text pointed to by persistent responses to it—for example, the impression that Milton's God in Paradise Lost speaks like a school divine-may be "simply ... the unintended and unavoidable negative conseauence of
I the artist's positive constructive intention" (in order to justify the ways of
1 God to men, Milton needed to have God speak, but it was inevitable that the
I transcendent Being would sound disappointingly human) (Rader 38).
I Still, such authorial limitations do not undermine our key point: once
i you decide to take a rhetorical perspective, the best way to make initial sense
of texts is to treat them as if they are intended to be made sense of—and then,
| once we've reconstructed that multidimensional sense, we can take the next
I step of evaluating the author's communication. Thus, when Huck heartlessly
answers Aunt Sally's question about whether anybody was hurt by the blown-
1 0L1t cylinder head—'"No'm. Killed a nigger'"—and when she replies equally
heartlessly—'"Well, it's lucky; because sometimes people do get hurt'"—as
rhetorical theorists we can reasonably assume that Twain designed those
I responses as something for us to interpret. He didn't want us to stop at their
* literal meaning but wanted us to go further, to decide what they reveal about each character, how those revelations influence our ethical judgments of
A them, and how the exchange affects our understanding of the larger narra-
1 live. Those decisions may not be easy—is Huck reverting to his old self? is
1 he simply playing a role for safety?—but they become either impossible or
I arbitrary if we don't assume that someone chose those words for a reason.
1 Once we make those decisions about the implied Twain's reasons, we are in
1 a good position to evaluate his choices, including the specific diction and
1 syntax of the exchange, his placement of it at this point in the progression,
i and, indeed, his decision to include it at all. In contrast, deciding to censure
f or defend Twain's choice before considering his purposes typically means
1 projecting our values onto Twain's novel—and thus forgoing our chance to
I understand its designs on us, much less to be influenced by them.
I While our most important commitment is to the role of authorial agency
1 in narrative communication, we also endorse the concept of the implied
l author that Booth introduced in The Rhetoric of Fiction. Booth defined the
1 implied author as the version of himself or herself whom the actual author
I constructs and who communicates through the myriad choices—conscious,
I intuitive, or even unconscious—that he or she makes in composing and
1 revising a narrative. The concept has subsequently been contested, with
I some theorists such as Gerard Genette—and our collaborator David Her-
* man—saying that it is unnecessary and others such as Mieke Bal saying that I it is too imprecise. At the same time, others have found it very valuable. We I side with the last group for a number of reasons:
I 1. It recognizes that writing narrative is inevitably an act of self-presenta-
I tion. Booth points out the often significant difference between the biographi-
) cal authors we learn about from reading their biosraohies (or even from
2: AUTHORS, NARRATORS, NARRATION
I
JAMES P H E LAN AND PETER J. RABINOWITZ
personal acquaintance) and the versions of those authors we come to know from encountering them in their narratives. Not all writers are self-conscious as they create their implied authors, but many—like Mark Twain—are. In his texts, Samuel Clemens constructed an image of himself (or more accurately a series of images of himself) that he believed deserved its own name. Strikingly, that name and the dominant image associated with it had effects that the biographical Clemens could not completely control: for example, many of his occasional political writings from the beginning of the twentieth century remained unpublished at the time he wrote them at least partly because they upset the image of the implied author of the popular texts.
The concept of the implied author also gives us purchase on a number of serious social and historical questions that are otherwise hard to navigate. Consider, for instance, the relations between gay writers and the images of themselves they construct when writing under conditions that make coming out dangerous. Consider, alternatively, the literature of self-justification, works by authors who are trying to present selves that counter widespread criticism of their actions (for instance, the multiple "authors" of the film On the Waterfront, some of whom were at least in part trying to present their decisions to testify before the House Un-American Activities Committee in a favorable light).
2. It gives us a useful way to talk about intention. A rhetorical approach is inevitably tied up with intention; yet, as many critics have rightly pointed out, the actual intentions of the actual author are often—even usually-unknowable. Furthermore, the actual author is not always the best judge of his design. Ernest Hemingway, when asked about an apparent contradiction in the attribution of dialogue to the two waiters in "A Clean, Well-Lighted Place," said that he thought the story was fine. But after his death, some Hemingway critics persuaded his publisher Scribners' that the implied Hemingway could not have intended the contradiction and so Scribners' agreed to amend the text (rightly in our view). More generally, the aim of the rhetorical approach is not to determine the conscious intention of the actual author (although, if available, that may be one piece of relevant information) but rather to discern the system of intentionality that explains why the text has this particular shape rather than some other one.
3. It helps explain why we often come to know different versions of the same actual author in different texts. The version of himself that Twain constructs in Adventures of Tom Sawyer is far less pessimistic than the version he constructs in "The Man That Corrupted Hadleyburg," much less "The United States of Lyncherdom." The attitudes toward women in the implied Ernest Hemingways "The Short, Happy Life of Francis Macomber"
are markedly different from the attitudes toward women in his "Hills Like White Elephants." In these cases, we have the same actual author but distinct implied authors.
4. It gives us a way to talk about texts with problematic authorship. This includes, for instance, ghostwritten, anonymous, and fraudulent texts. It also , includes collaboratively written texts (such as the one were writing here): the
1 two (or more) actual authors construct a hybrid version of their actual selves,
and it is that hybrid version that readers come to know.
We want to stress that in our view the implied author, though knowable through the text, is not a textual construct equivalent to one of the characters but rather the agent who constructs the text. In other words, in the light of our main commitment, to authorial agency, we are less invested in the distinction between actual and implied author than we are in the view that texts . are not collections of free-floating signifiers but purposive communicative
f actions designed by some authorial agent. At the same time, as we noted in
our introduction, we are interested in analytic categories only to the extent that they have some significant payoff—and we recognize that in many critical discussions the key distinction is between authorial agency and textual free play rather than between the actual and the implied author. In those cases, we will often use the name of the author to cover both the real author and the implied author.1
I
NARRATORS
One of the most important choices an author makes is about the kind of narrator to employ. Working in a specific context—in particular, trying to counter the prevalent orthodoxy that placed "showing" in a position over "telling," and trying to get beyond the easy, but misleading, distinction between first-and third-person narration—Booth was especially concerned with the distinction between dramatized and undramatized narrators. His most durable contribution, though, was his development of the concept of the "unreliable narrator."
Booths initial conception was relatively simple: a narrator is "reliable when he or she speaks for or acts in accordance with the norms of the work (which is to say, the implied author's norms), unreliable when he does not" (158-59). Since then, however, rhetorical narrative theorists have introduced a number of refinements. For instance, Jim has developed a distinction between restricted narration and unreliable narration2 and a taxonomy of six types of unreliability. Both developments arise from his observation
2: AUTHORS, NARRATORS, NARRATION
that narrators perform three primary tasks: they report (along the axis of facts, characters, and events), interpret (along the axis of knowledge or perception), and evaluate (along the axis of ethics). When a narrator performs only one of the three tasks and the author uses that restriction to communicate something that the narrator is unaware of, the implied author is using restricted narration. In the most common kind of restricted narration, the implied author will have a naive narrator reliably report the events but not attempt to interpret or evaluate them, and in so doing communicate an interpretation or evaluation that is beyond the capacity of the naive narrator. Here's an example from Huckleberry Finn: "Tom Sawyer, he hunted me up, and said he was going to start a band of robbers, and I might join if I would go back to the widow and be respectable. So I went back" (32). Twain restricts Huck to the task of reliably reporting events, thereby implying that Huck fails to grasp the irony of what he reports: the contradiction in Toms logic that allows him to use respectability as an eligibility condition for joining a band of robbers.
When a narrator performs any of the three tasks inadequately (as measured against how the implied author would perform it), then we have unreliable narration. But narrators can be inadequate in two main ways: by distorting things or by failing to go far enough. Consequently, narrators can be unreliable by misreporting, misinterpreting, and misevaluating (in these cases readers need to reject the narrators version and, if possible, replace it with another one) and by underreporting, underreading, and underevaluat-ing (in these cases readers need to supplement the narrator's version).
Just as significant is Jim's more recent distinction between estranging unreliability ("unreliable narration that underlines or increases the distance between the narrator and the reader") and bonding unreliability ("unreliable narration that reduces the distance between the narrator and the reader") (Phelan, "Estranging" 222-23). In the wake of Booths work, the default position was that unreliability served to distance us from the narrator—to be an unreliable narrator was, ipso facto, to be somehow deficient (in particular, given Booth's ethical emphases, morally deficient). But in fact our relations to unreliable narrators, like our relations to other people, are more complex than this kind of judgmental clarity would suggest, and authors use this complexity to create a range of emotional and ethical effects. The bonding/estranging distinction recognizes that with unreliable narration, this range extends from affective and ethical repulsion at one end to affective sympathy and ethical admiration at the other. Booth's original conception privileged the relationship of reader and implied author: "the emotions and judgments of the implied author are ... the very stuff out of which great fic-
(AMES PHELAN AND PETER J. RABINOWITZ
tion is made" (86)—and Booth eloquently demonstrated how sharing those
Iemotions and judgments often involves our communion with the implied author either through her surrogate, the reliable narrator (as in Jane Austens Emma) or "behind [the] back" (Rhetoric 159) of the unreliable narrator (as in Ring Lardner's "Haircut"). Recognizing bonding effects illuminates the importance of a different kind of communion accompanying the behind-
Ithe-back connection of unreliability: a communion between readers and a narrator who could never be mistaken for the implied authors surrogate. In this way, the bonding/estranging distinction adds an important dimension to our understanding of the workings of narrative communication.
We can clarify these theoretical points by exploring the varying nature !f and effects of unreliable narration in Huckleberry Finn. First, a fairly simple
1 example: Huck describes the widow Douglass response to his return to her
* home this way: "The widow she cried over me, and called me a poor lost
lamb, and she called me a lot of other names, too, but, she never meant no 1 harm by it" (32). This passage uses a combination of reliable reporting, misin-
1| terpreting, and underevaluating in the service of bonding effects. Huck mis-
|| interprets the widows joyous religious references as name-calling because he
j doesn't recognize the New Testament source—and that misinterpreting leads
him to undervalue the ethical quality of her response. Yet this comic failure of understanding simultaneously reveals a moral strength. Although Huck's ignorance means that he fails to grasp both the extent of the widow's joy and f her beliefs about what his return means, Twain demonstrates that Huck's
g ethical compass is sufficiently sensitive for him to appreciate that she "never
meant no harm." Again, the overall effects are to bring us affectively and ethically closer to Huck even as we continue to register our interpretive distance from him. Simply to call him an "unreliable narrator" fails to capture these effects and, consequently, is itself a kind of underinterpreting. i The same kind of interplay—with more subtlety and greater conse-
| quences—marks Huck's self-examination in Chapter XXXI, which leads to
his famous decision to go to hell rather than inform Miss Watson of Jims whereabouts.
II know very well why [the words of my prayer] wouldn't come. It was because my heart warn't right; it was because I warn't square; it was because I was playing double. I was letting on to give up sin, but away inside me I was holding on to the biggest one of all. I was trying to make my mouth -J say I would do the right thing and the clean thing, and go and write to that
| niggers owner and tell where he was; but deep down in me I knowed it
| was a lie—and He knowed it. You cant pray a lie—I found that out. (200)
2: AUTHORS, NARRATORS, NARRATION
JAMES P H E L A N AND PETER ). RABINOWITZ
the one discussed earlier, Twain is now asking us to work through something far deeper. On the one hand, there's still a level of broad comedy, since, for Twain and his authorial audience, Huck's fear of going to hell is just as much a superstition as Jim's fear of taking up a snakeskin in his hand. At the same time, we transcend Huck's misinterpretation and admire the courage that his choice reveals. While Huck is sure that he's damned, Twain's audience is even more certain that he is making the right ethical choice.
As our tone suggests, we greatly admire Twain's handling of the relationships among author, narrator, and audience in this passage, and we extend our admiration for his craft to the whole first two-thirds of the narrative. In chapter 7, which explicitly takes up the question of evaluation, we will discuss what we regard as Twain's far less successful narration in the last third of the novel, the infamous Evasion section.
Here we have a case of misevaluating being used for bonding effects. Huck judges himself to be ethically deficient, while the implied Twain guides us to judge Hucks inability to act against Jim's interest as a sign that he is acting according to a higher ethical standard.
Huck then writes the letter to Miss Watson, revealing Jim's location, so that he'll be able to pray for help to stop sinning. He immediately feels better; but before he prays, he starts thinking about the River trip:
And I got to thinking over our trip down the River; and I see Jim before me, all the time, in the day, and in the night-time, sometimes moonlight, sometimes storms, and we a floating along, talking, and singing, and laughing. But somehow I couldn't seem to strike no places to harden me against him, but only the other kind.... and at last I struck the time I saved him by telling the men we had small-pox aboard, and he was so grateful, and said I was the best friend old Jim ever had in the world, and the only one he's got now; and then I happened to see to look around, and see that paper. (200-201)
At this moment, Huck is reliable as reporter, interpreter, and evaluator—and this burst of insight serves to reinforce the previous bonding unreliability. \ It does so by revealing the ethical heart of Hucks previous misevaluating, showing us that it applies not to his judgments of Jim but to his conception of "the right thing and the clean thing" in this situation. To send the letter * would be to follow the dictates of conventional Christianity, but it would simultaneously follow the dictates of a slave society that has used Christianity as one of its buttresses. Huck intuitively recognizes that by following both sets of dictates, he would be going against everything his experience of Jim has taught him about the right and clean thing to do. Thus, when Huck makes his decision to tear up that paper and—in a return to unreliability-— evaluates his action most negatively, we feel our strongest sympathy and our greatest ethical approval of his actions.
"All right, then, I'll go to hell"—and tore it up.
It was awful thoughts, and awful words, but they was said. And I let them stay said; and never thought no more about reforming. I shoved the whole thing out of my head, and said I would take up wickedness again, which was in my line, being brung up to it, and the other warnt. (201)
With this climactic misevaluation, the bonding unreliability reaches its apex. In particular, having trained us with comic passages about religion, including
CONVERSATION AS NARRATION
One consequence of viewing narrative primarily as a communicative event rather than primarily a textual structure is that we are less beholden to the distinction between story and discourse than many other theorists. That distinction neatly separates elements of story (events and existents, including settings and characters) from elements of discourse (structure and modes of narration) in the interests of analytical precision. In this view, scenes of character-character dialogue are part of story, since they are events involving characters, while any narratorial commentary on the dialogue is part of discourse. But from the perspective of rhetorical theory, such analytical precision comes at the cost of explanatory power, since authors can use character-character dialogue both as events and as modes of narration. Thus, scenes of character-character dialogue with interspersed narratorial commentary often show an author using different resources of narrative (characters, the narrator, the occasion of the dialogue) in order to combine the representation of an event with reporting, interpreting, and evaluating.
At the end of Chapter XII, Twain tells about a climactic moment in Huck and Jim's experience on board the wrecked steamboat The Walter Scott. Huck has been reporting the conversation he overhears between Jake Packard and Bill about whether they should kill Jim Turner, which ends with Packard convincing Bill that they can leave the wreck by using its small boat and then watch to make sure he drowns when the wreck breaks up. That dialogue itself is obviously full of reporting, interpreting, and evaluating, but even more salient are the narrative functions of the dialogue between Huck and Jim
Robyn Warhol
"Quick, Jim, it ain't no time for fooling around and moaning; there's a gang of murderers in yonder, and if we don't hunt up their boat and set her drifting down the river so these fellows can't get away from the wreck, there's one of em going to be in a bad fix. But if we find their boat we can put all of em in a bad fix—for the Sheriff'11 get em. Quick—hurry! I'll hunt the labbard side, you hunt the stabbard. You start at the raft, and—
"Oh, my lordy, lordy! Raf7. Dey ain no raf' no mo,' she done broke loose en gone!—en here we is!" (86-87)
In Huck's portion of this dialogue, Twain shows him offering to Jim—and by extension to Twain's actual audience—ethical judgments about "these fellows" as well as interpretive judgments about what he and Jim need to do. These judgments are as clear and significant as they would be if Twain had switched the mode to Huck's telling his narratee about how he'd sized up the situation. However, by conveying Huck's judgments in his address to Jim, Twain also tells us about the way Huck implicitly trusts Jim even as he shows Huck's instinctively clear ethical judgments and his shrewd strategy for dealing with the situation. Jim's line in turn reports a significant new event, which is itself a serious local complication of the instabilities, along with his affective response. Twain also invites us to infer the significance of what Jim does not say, namely, "I told you so," since Jim had advised against leaving the raft to board The Walter Scott. The larger point here is that Twain is using the resources of character-character dialogue as narration by other means.3
Who is speaking? To whom? In what circumstances? These most basic questions of classical narrative theory provide the starting place for a feminist narratological study of narration. For a gender-centered analysis, the personification of narrator and narratee implied in the usage "who?" and "to whom?" is entirely appropriate, because feminist narratology so often treats a text's situation of enunciation as if it were an exchange between embodied persons. "Who is speaking?" refers not just to the narrator(s) but also to the author; "To whom?" asks about actual audience, as well as narratee. And for the feminist narratologist "In what circumstances?" means "At what moment in history?" as well as "At which diegetic level?" or "With what degree of localization?" A text has its origin in the material world, a world where gender shapes perceptions and realities that go into the writing and reading of books. Therefore the identity, experience, and socio-cultural-historical circumstances of the author—not to mention the reader—are important in understanding the ways that narrative participates in the politics of gender.
In this belief, feminist narratology departs radically from its classical roots. For Gerard Genette, Gerald Prince, and Dorrit Cohn—the three most influential of the theorists whose narratological models served as the jump-ing-off point for the first generation of feminist narratologists—narrative discourse is an abstract structure, to be rigorously distinguished from its real-world creator or receivers. Feminist critics from Virginia Woolf forward have regarded abstractions suspiciously, however, constantly aware of the gendered body that writes or reads every text. For this reason, the person who wrote the text is a living presence in feminist narratology. Too deeply invested in poststructuralism to place much value on "authorial intention," feminist narratology nevertheless asks how the authors gendered experiences of a particular time in a particular place affect the structures he or she employs in putting together the story and discourse that comprise narrative. Unlike classical narratologists, critics who use feminist narrative theory tend to be as interested in the interpretation and evaluation of texts as in the description of how texts work. In asking "What does this narrative mean?" or "Is this text any good—in terms of aesthetics, ethics, or politics?" the feminist narratologist is not barred, as are her classical forebears, from consulting what is known about the Author herself.
Jane Austens history, then, as a genteel clergyman's daughter writing in England during the early nineteenth century colors the feminist narratologists analysis of Persuasion. Because she has become a cult figure both inside and outside academia, Austen is particularly present in most analyses of her
2: AUTHORS, NARRATORS, NARRATION
ROBYN WARHOL
work. Often, author and narrator are elided in what is understood as "Austens voice," a distinctive blend of ironic hyperbole, orderly syntax, parallel phrasing, free indirect discourse, and periodic sentences. From the beginning many critics have tended to regard that voice as a representation of Austen herself. Departing from the sentimentalized Victorian image of "gentle Jane" disseminated by her nephew J. E. Austen-Leigh's biography of Austen, and taking a cue from the mid-twentieth-century belief, articulated by D. W. Harding, that Austens voice speaks with "regulated hatred" about characters "she herself detests and fears" (12), many feminist critics identify the novelist's persona with Cassandra Austen's one surviving sketch of her sister: critical, quizzical, sharp, humorous, impatient, penetrating, active.4 Austen's evident authorial investment in the conventional marriage plot posed a stumbling block for some second-wave feminist critics, but the kind of gender-centered narrative criticism I practice finds Austen's feminist message less in the details of her stories than in her handling of narration.5
Austen's famous irony provides a classic example of what feminist narra-tologists have called "double-voicing," in that her narrator characteristically makes utterances that appear to mean something other than precisely what is said. Thus, one voice speaks the literal meaning of the utterance, and a second, implicit voice ironizes that literal meaning. But sometimes the narrator will set up the double-voicing by straightforwardly describing characters or situations as the narrator of Persuasion does in this early pronouncement on Anne's father: "Vanity was the beginning and the end of Sir Walter Elliots character; vanity of person and of situation" (10). In the beginning of Persuasion, as in her other novels, Austen's narrator lays out the basic characteristics as well as the familial and financial situations of all her principal characters, thus setting up the range of possibilities for ways the story might develop. Having made so clear a statement of "the beginning and end" of Sir Walter's character, the narrator thereafter freely exercises her penchant for explaining Sir Walter's thoughts and actions as though they were perfectly sensible, while silently communicating the ludicrousness of what the character says and does. For instance, once Lady Russell has persuaded Sir Walter that his straitened financial circumstances require him to rent out Kellynch Hall, the narrator deadpans her description of his attitude, a deadpan that depends on our recognizing her mingling of Sir Walter's voice with the narrator's in free indirect discourse:
Sir Walter could not have borne the degradation of being known to design letting his house.—Mr. Shepherd had once mentioned the word "advertise;"—but never dared approach it again; Sir Walter spurned the idea of its
being offered in any manner; forbad the slightest hint being dropped of his having such an intention; and it was only on the supposition of his being spontaneously solicited by some most unexceptionable applicant, on his own terms, and as a great favour, that he would let it at all. (19)
Austen's narrator thus exposes Sir Walter's silliness without directly criticizing it. One way to think about how her irony works is to realize that if Sir Walter could read this passage, he would recognize what the narrator is overtly saying as "quite right." This realization, in turn, opens up the presence of Sir Walter's voice and the beliefs associated with it in this passage, especially in "degradation," "spurned," "forbad," "most unexceptionable applicant" and "great favour." Austen's narrator—or rather the implied author informing that figure—counts on the authorial audience to recognize the vainglorious overstatement in Sir Walter's language, especially when juxtaposed with such plain speaking as "advertise," and thus get the joke. This is classic feminist "double-voicing": in this paragraph Austen's narrator takes the patriarch down without uttering a single word against him.
As this passage indicates, Austen's ironic, double-voiced narration is intimately tied to her trademark use of free indirect discourse. Moving seamlessly between the narrator's opinions and the characters' thoughts or utterances, the narrator uses the characters' own words against them. In Persuasion Austen uses this technique not only when representing characters who are being satirized—Sir Walter, Lady Russell, Anne's sisters Elizabeth and Mary, the Musgrove girls, Mrs. Clay, Mr. William Walter Elliot, to mention just a few—but also when representing the heroine herself. Frequently the narrator renders Anne's thoughts as Anne might have voiced them had she spoken, seldom shielding the heroine from the narrator's usual ironic practice. For example, in one of the scenes where Anne anticipates seeing Frederick Wentworth eight years after the rupture of their love affair, the narrator writes, "What was it to her, if Frederick Wentworth were only half a mile distant, making himself agreeable to others!" (51). The exclamation point expresses the energy Anne must exert to convince herself that Went-worth's whereabouts are nothing to her, at the same time that it indicates the narrator's ironic take on Anne's thoughts. When Anne hears from her sister Mary what Wentworth has said upon seeing Anne again, the free indirect discourse with which the narrator renders the heroine's thoughts again gives the heroine the lie:
"So altered that he should not have known her again!" These were words which could not but dwell with her. Yet she soon began to rejoice that she
2: AUTHORS, NARRATORS, NARRATION
had heard them. They were of sobering tendency; they allayed agitation; they composed, and consequently must make her happier." (54)
Succeeding events—not to mention the state of anxiety that persists for her throughout the rest of the novel—show the momentary self-delusion Anne indulges during this burst of rationalization. Even this most admirable of Austens heroines is not immune from the narrators ironic double-voiced discourse.
Austens narrators speak from the position that used to be called by the oxymoronic term "limited omniscience," more properly described as het-erodiegetic narration focalized almost exclusively through the protagonist. What the narrator "knows" is a question I will take up later, when discussing perspective below; who the narrator "is" can be understood as a separate question. For the most part a disembodied voice, Austens narrator only rarely uses the first person, most often near the end of her novels. Persuasion is no exception, as the narrators rare "I" appears in the first paragraph of the novel's final chapter. Because Austens narrator is never a character within the diegetic world, the brief appearances of her "I" make for moments of metalepsis, reminders that the diegesis hangs by the thread of the narrators words and, in turn, by the novelist's invention. Significantly, these metalepses frequently coincide, as in Persuasion, with the much-longed-for resolution of the marriage plot. Just when she comes to an opportunity to give juicy details about the heroine's blissful union with the hero, Austen's narrator pops up with a self-reference, explicitly refusing, in Austen's characteristic way, to narrate those very details. In Persuasion, the narrator's "I" appears in a moment of unnarration, relegating romantic details to rhetorical questions and declining to draw a conventional moral:
Who can be in doubt of what followed? When any two young people take it into their heads to marry, they are pretty sure by perseverance to carry their point, be they ever so poor, or ever so imprudent, or ever so little likely to be necessary to each others ultimate comfort. This may be bad morality to conclude with, but I believe it to be truth; and if such parties succeed, how should a Captain Wentworth and an Anne Elliot, with the advantage of maturity of mind, consciousness of right, and one independent fortune between them, fail of bearing down every opposition? (199)
ROBYN WARHOL
lished posthumously together with Persuasion, comes to a similarly open conclusion:
To begin perfect happiness at the respective ages of twenty-six and eighteen is to do pretty well; and professing myself moreover convinced that [his father's] unjust interference, so far from being really injurious to their felicity, was perhaps rather conducive to it, by improving their knowledge of each other, and adding strength to their attachment, I leave it to be settled, by whomsoever it may concern, whether the tendency of this work be altogether to recommend parental tyranny, or reward filial disobedience. (240)
"Professing [her] self convinced" that Henry and Catherine's relationship has benefited from adversity, the narrator refers to them as though the characters inhabited the same plane as the narrator. Referring in the same sentence to their status as creatures of a narrative creation (by mentioning "this work"), she immediately undermines the suggestion that Henry and Catherine have any substance beyond the fictitious world. That the narrator steps into the diegetic frame to draw attention to herself by making jokes about the morals of her marriage-plot conclusions should come as a surprise. These metaleptic moments are serving a feminist purpose, destabilizing what is supposed to be the conventional resolution of the happy ending. What is the moral of the story, or what should the moral of the story be? Austens narrator isn't telling.
The moral of a story is attributable to the author, as Austen's narrator is well aware; and Persuasion is not the only one of Austens novels that ends on a moral the narrator declares to be questionable. Northanger Abbey, pub-
DAVID HERMAN
David Herman
In this chapter I focus on narration as a mode of situated communicative action, which can be analyzed, on the basis of textual performances, in the same way that other sorts of action are—namely, in terms of reasons for acting. We can think of a nested structure of actions in this connection: local textual choices subserve the more global purposes of narrative worldmaking, which are nested in turn in a still broader ecology of representational goals. In everyday discourse I can recruit from words, gestures, and other communicative resources to construct a narrative that I use in turn to account for (or repudiate) someone's conduct, for example, or to justify the attachment I feel for a particular place. Likewise, written, literary narratives like McEwans can be characterized as a form of communicative action, for the uptake of which inferences about authors' (and sometimes narrators') reasons for acting are not only pertinent but necessary.
NARRATIONAL ACTS AND REASONS FOR THEM
What distinguishes actions from mere behaviors is that whereas both are the effects of causes that can be described in physical or material terms, only in the case of actions does it make sense to ask, too, about reasons—that is, about why an agent has chosen to act in a particular way instead of in other possible ways (Brockmeier; Malle).6 For example, anatomical structures and physiological mechanisms provide the causal basis for my ability to extend my index finger outward while curling my other fingers in against the palm of my hand. But when it is situated in the frame of (communicative) action, this bodily movement or behavior can also be parsed as a pointing gesture, which I perform for reasons that you provisionally ascribe to me as part of the process of interpreting my behavior as a gesture, rather than an uncontrolled reflex or unintentional tic. Reasons for acting can be analyzed, in turn, into interlocking sets of beliefs and desires (or, more generally, prepositional and motivational attitudes), as when you interpret my bodily movement as a pointing gesture by ascribing to me the belief that there is something in our shared environment that is worth calling to your attention, as well as the desire to elicit your attention by pointing at the object in question.7
Narrational acts do stem in part from causes—for example, the movement of vocal cords or of fingers on a computer keyboard. But beyond this they result in textual performances on the basis of which reasons for acting—reasons for producing a narrative that has a particular plot structure,
3j
mode of temporal sequencing, thematic focus, etc.—can be ascribed to those who engage in narrational conduct. Narration, in other words, does involve behavior that is explicable in causal terms; yet it cannot be exhaustively described as behavior but rather falls in the subcategory of behaviors that constitute actions and that are engaged in for reasons. Interpreters impute these reasons to narrating agents to make sense of their behaviors as communicative actions in the first place.
Thus, though it is of course possible to explore what physical mechanisms and processes may have caused a writer's behavior, when one moves to the domain of narrational action why-questions, or questions about reasons, lv*. i >me pertinent. Why does McEwan use analepses or flashbacks to Flor-ciiti's and Edward's courtship and earlier family histories, instead of a more straightforwardly chronological presentation of events? Likewise, what is the reason for the text's periodic shifts between Florences and Edwards vantage points over the course of the story, when in principle McEwan could have used a distanced, authorial mode of narration throughout or, alternatively, a single reflector or center of consciousness for the duration of the novel (Stanzel)? In connection with the novel's non-chronological narration, to be discussed in more detail in chapter 3,1 impute to McEwan the belief that (in the context of this fictional world) Florence and Edward act the way they do in the present moment in part because of how they have been shaped by past situations and events; I also impute to the author the desire to foreground the impact of those earlier contexts on the newlyweds' current choices and experiences. Similarly, to anticipate issues discussed more fully in chapter 4, in connection with the text's use of what Genette would term "variable focal-ization" (Narrative Discourse 191), or the adoption of shifting perspectives on the world of the narrative, I impute to McEwan the aim of underscoring the very different profiles that events can assume depending on one's background assumptions, life history, and gender identity—as well as the importance of facing such differences squarely and negotiating them carefully.
Clearly, these are only initial, rough characterizations of the sorts of reasons for narrational acts that interpreters might ascribe to McEwan on the basis of the text. A fuller account would need to specify more precisely the procedures for ascription, the textual designs that motivate those ascriptions, and the pairings of prepositional attitudes (believing, doubting, or imagining x) and motivational attitudes (desiring, loathing, or feeling indifferent toward y) that help constitute the reasons being ascribed. Along these lines, the technical concepts and nomenclatures developed by theorists of narrative can be viewed as frameworks for categorizing and differentiating among textual features in ways that bring to light their functions as story-
DAVID HERMAN
AGENTS OF NARRATION
Up to now, I have focused on the process of narration rather than on the agent or agents who carry it out, or the (more or less direct or oblique) inferential pathways that lead from acts of narration to performers of those acts. Granted, along with any previous familiarity that a reader may have with a given writer's work, paratextual features of novels like McEwans—for example, a title page that includes the phrase "A Novel" under the main title-already help channel and delimit inferences about the reasons for acting that can be assumed to underpin specific textual choices in the narrative. But though paratextual cues and prior knowledge help mold ones general or global approach to narrational acts, making sense of those acts on a local, moment-by-moment basis entails grounding them in reasons for acting, which in turn entails moving along an inferential route that points in the opposite direction—from the act to the agent of narration. To sketch out how one might begin describing this pathway and some of the implications of pursuing it, I turn now to broader issues that arise when one approaches the telling of stories as a form of communicative action whose uptake requires reasoning about tellers' reasons.
Two key concerns can be singled out. For one thing, a focus on narration as action exposes the limitations of anti-intentionalism in narrative study, bringing to the fore how ascriptions of communicative intentions lie at the foundation of narrative experiences. But to what agent or agents can (or should) these intentions be ascribed? Here emerges a second set of issues, concerning the notions of author and narrator and the relation between them. In an approach that openly acknowledges the need to ascribe intentions to narrating agents, I argue, authors or story creators become centrally important, while the category of "narrator" will be more or less
ii-fgest that an approach to narration as action removes the basis for appeals to an implied author as the source of intentions projected in or assumed by -sarrative.
1. Reclaiming intentions: The approach to narration being sketched here develops further the critique of anti-intentionalist assumptions that I outlined in a previous study (Herman, "Intentional"). Characterizing narration as a form of communicative action whose interpretation involves—indeed, requires—ascriptions of reasons for acting is a manifestly intentionalist line of inquiry.8 Hence my approach contrasts starkly with W. K. Wimsatt and Monroe Beardsley's discipline-shaping essay "The Intentional Fallacy" (originally published in 1946), which argued that using an author's intentions as a yardstick for literary interpretation is neither possible nor desirable. More specifically, Wimsatt and Beardsley argued against confusing the author's iltsigning intellect, which they admit to be the cause of the poem, with a -Mndard for judging or interpreting the poem; this is the broader "genetic fallacy" (confusing the nature of a thing with its origin) of which for Wimsatt and Beardsley the intentional fallacy is a special case. As the previous discussion indicates, however, origins can be subdivided into reasons and causes, depending on whether one is examining actions or behaviors, respectively. To reiterate: construing people's doings in terms of reasons for acting is the basis for understanding those doings as actions in the first place.
Thus Wimsatt and Beardsley begged the question, arguably, when they equated (a) an interpretive practice based on ascriptions of intention with (b) commission of the genetic fallacy. Interpreting an act of narration by ascribing reasons for it is no more a confusion of the act with its causal origins than is interpreting someone's rude behavior by appealing to reasons why he or she is acting that way. To the contrary, it is indeed part of the nature of an action for it to be explicable through an account of how it arises from or originates in a reason (or set of reasons) that involves intentions and other motivational and propositional attitudes. In this same connection, note that Booth developed the concept of the implied author to be able to factor in readers' inferences about the communicative intentions and purposes manifested in narrative texts, but at the same time avoid running afoul of the anti-intentionalist strictures set out by Wimsatt and Beardsley. The better solution, I argue in my next subsection, is to dispute the anti-intentionalists premises from the start, anchoring the process of narrative interpretation in defeasible or possibly wrong inferences concerning authors', not implied authors', communicative intentions.
2. Authors and narrators: But shouldn't reasons for telling be ascribed to narrators rather than authors? There is no one-size-fits-all answer to this
2: AUTHORS, NARRATORS, NARRATION
DAVID HERMAN
fests as it unfolds in a particular context of storytelling and story interpretation, determines the nature of the inferential pathway that leads back to the agent or agents who performs/perform it. The pathway may be relatively ^ direct, as in (some) autobiographical narration, or relatively indirect, as in first-person or homodiegetic narratives that feature unreliable tellers—with an intermediate position being occupied by texts that, like McEwan's, are f narrated heterodiegetically. The pertinence of the concept of "narrator" varies across these and other storytelling situations. In On Chesil Beach, the narration is conducted in a way that suggests the presence of a teller arranging "I the sequence in which events are recounted, filtered through Edwards and Florence's perspectives, and so on. But this teller remains nonpersonified, uncharacterized, all but obviating the need to draw inferences concerning the teller's (in contradistinction to the author's) reasons for narrating events in the way they are presented (compare Walsh 69-85). i
By contrast, McEwan's 1978 novel The Cement Garden is recounted by Jack, one of the family members whose responses to the death of their parents constitute the subject of the narrative. Here two levels of narrational action (and the reasons that respectively underpin them) need to be considered: that by means of which the author evokes Jack as a character who tells about his efforts to bury his dead mother in concrete, his incestuous relation- t ship with his sister, and so forth; and that by means of which Jack produces the narrative that details these same events, with Jack's own reasons for tell- ^ ing his story in the way that he tells it now coming into question. Drawing % inferences about the motivational and propositional attitudes that account for Jack's narration thus requires navigating a nested structure of reasons * within reasons, whereby readers frame inferences about Jack's intentions and other attitudes by grounding them in inferences about McEwan's reasons for creating Jack as a fictional person who has such attitudes.9 *
This nested structure assumes a different shape in the 1993 film version of The Cement Garden. In this case there is a diversified set of narrating agents, including cinematographers, the film's director (Andrew Birkin), and * its editors, among other story creators, who collaborate on the production of what viewers construe as a more or less unified narrational action (Bor-dwell). In this complex, collaborative act of narration the characters' own reasons for acting once again figure, at the diegetic level, as key targets for interpretation. But arguably the inferential processes activated by the film, like those triggered by a ghostwritten text or a hoax such as James Frey's A Million Little Pieces, are of the same basic kind as the inferences prompted by third-person "authorial" narration or first-person narration, whether reliable or unreliable. Thus, whereas Genette in using Occam's razor to critique the
concept of the implied author made an exception for ghostwritten narratives and hoaxes {Revisited 146-47), I suggest that in these cases too an inten-tionalist approach grounded in the logic of narrative worldmaking warrants greater parsimony.
Here I target for critique ideas associated with what has become known as the narrative communication diagram, which features not only authors and readers but also implied authors and implied readers, as well as narrators and narratees. My argument is that this diagram, precisely by hedging against intentionalism, opens the door to what Alfred North Whitehead (58-59) once described as the problem of "misplaced concreteness." This problem comes into play when the heuristic status of models for inquiry becomes obscured or forgotten. In direct proportion with the force of its ban against ascriptions of intentions to authors, the diagram invests the construct of implied author with operative power, that is, knowable reasons for acting; and an entity originally designed to serve heuristic purposes thus becomes reified as the basis for interpretations of a text. But an openly intentionalist approach to worldmaking avoids such misplaced concreteness. From the perspective afforded by this approach, hoaxes and collaboratively authored texts (like other narrational acts) cue interpreters to frame defeasible or possibly wrong inferences about story creators' reasons for narrating. Accordingly, rather than distinguishing between the actual Frey and an implied author projected by his text, I can more parsimoniously characterize the text as one that, like all other narratives, was designed as a blueprint for world building. If I do not know about the controversy surrounding the text and Frey's eventual confession that he fabricated key elements of his narrative, I will assume that the storyworld evoked by the text is capable of being dis-confirmed through triangulation with other accounts of the events at issue. Once I know that James Frey concocted incidents in his putatively nonfic-tional account, however, I will use different strategies for world-construction, modifying the inferences I initially drew concerning Frey's reasons for narrating in the way that he did the events associated with his addiction and recovery. Specifically, I will interpret the text as one purposely—intentionally—designed by its author to prompt infelicitous modes of world-building; in other words, I will factor this additional assumption into the broader context that allows me to situate Frey's narrating act and explicate it in terms of reasons. And the same goes for collaboratively created narratives like movies. Once I have been schooled in how movies are actually made, I will be able to situate Birkin's version of The Cement Garden in a different context. I can now see the film as the result of a sustained effort on the part of the movie's multiple creators to coordinate local feats of story-elaboration,
2: AUTHORS, NARRATORS, NARRATION
cobbled together more or less seamlessly, to produce the effect of a single act of narration.
I am advancing, then, a two-part argument against models based on the notion of the implied author. The first part of the argument is that the idea of the implied author arises from efforts to accommodate an anti-intentionalist position that I believe it is preferable to dispute from the start, by grounding the worldmaking process in defeasible ascriptions, to authors or story creators, of reasons for performing acts of narration. The second, related part of the argument is that talk of implied authors entails a reification or hypostati-zation of what is better characterized as a stage in an inferential process (see also my contributions to chapter 6 and to Part Two of this volume). At issue is the process by which interpreters, by ascribing to story creators reasons for performing acts of narration, build and rebuild narrative worlds whose contours may change in light of ongoing attempts to factor in textual as well as contextual information. If, for instance, it emerged that the details of Edward Mayhew's life history exactly paralleled McEwan's own (in a kind of reverse-Frey scenario), then I would have to alter my understanding of the relation between the author and the narrator, the actual world and the storyworld. But instead of trying to account for this change by multiplying explanatory entities, and by distinguishing between an author implied in or by the text and an author who actually created it, in my approach I assume that there are only authors or story creators whose reasons for producing a given narrative I may well need to recontextualize and thus interpret differently over time.
In turn—to anticipate issues discussed in chapter 6—an intentionalist approach to worldmaking has major consequences for accounts of the role of the reader in narrative contexts. Suffice it to say for now that the approach I sketch out in this volume is premised on an intentionalism without implied authors—and thus without implied readers. In place of the narrative communication diagram, my approach to narrative world-building grows out of a rigorously intentionalist but more minimal explanatory model that I abbreviate as capa. This model encompasses Contexts for interpretation (including contexts afforded by knowledge about narrative genres, an author's previous works, or Frey-like confessions); storytelling Actions performed within those contexts, and resulting in texts that function as blueprints for worldmaking; Persons10 who perform acts of telling as well as acts of interpretation; and defeasible Ascriptions of communicative and other intentions to performers of narrative acts—given the contexts in which the persons at issue perform those acts and the structure that their resulting narratives take on. I provide more details about the capa model, and its advantages over the narrative communication diagram, in my contribution to Part Two.
Brian Richardson
I w ill start with a standard account of author-narrator relations and then go on to see how antimimetic writers play with these boundaries. The author is the human being who actually puts pen to paper (or fingers to the keyboard) and puts his or her royalties in the bank; the implied author is the figure or sensibility we imagine the author to be, based primarily on our reading of the work; and the narrator is the fictional person (or persons) who is responsible for transmitting the narrative. Or, in the useful formulation of William Nelles: "The historical author writes, the implied author means, and the narrator speaks" (22). As the classic example has it, in Huckleberry Finn, the author is Samuel Clemens, the implied author or sensibility is that of "Mark Twain," and the narrator who relates the story is Huck Finn. The concept of an implied author is especially important when discussing co-written, ghostwritten, or anonymous works: political speechwriters all want to sound like the candidate who will speak their words; the multiple authors of a religious work, modern novel, or Hollywood movie want the material to sound as if it came from the same person; and the concept is essential when we ascribe anonymous works to a historical author or construct a single figure on the basis of a style that is repeated (e.g., "the 'Pearl' Poet" or "the Wakefield Master"). Two historical authors may also combine to provide a single authorial voice, as Conrad and Ford do in Romance. The concept of the narrator has been important historically for correctly dissuading readers from simplisti-cally equating the author of a text with its speaker: T. S. Eliot is not J. Alfred Prufrock, and Vladimir Nabokov is not to be equated with the narrator of Lolita: "my creature Humbert is a foreigner and an anarchist, and there are many things, besides nymphets, in which I disagree with him," explained the author (317). It is also possible for us to detect the views of an actual author in the mouth of a narrator or other character, including an unlikely one. I call this "transparent narration," a more direct and even nonillusionistic type of narratorial communication than what James Phelan has called "mask narration" (Living 201-4). For Phelan, mask narration occurs when an author uses a character narrator to express the author's beliefs; the masking can be subtly or sloppily done. By contrast, transparent narration breaks the frame of fic-tionality by making nonfictional statements in the narrative. There is no pretense of hiding; the mask is simply discarded as the line between author and narrator is erased. Thus, in Kurt Vonnegut s Slaughterhouse Five, the narrator identifies himself as a character in the novel's storyworld and simultaneously affirms the reality of the novel's historical testimony of the Allied firebomb-ing of Dresden: "That was me. That was the author of this book" (125).
2: AUTHORS, NARRATORS, NARRATION
BRIAN RICHARDSON
The concept of an implied author is particularly useful for discussing the variously unreliable narrators of modernist fiction; a fallible narrator presupposes a cunning implied author who wants the intended reader to perceive the strategy of unreliability and thereby "get the joke." Postmodern and other antimimetic authors, however, delight in collapsing established categories, and the triad of author, implied author, and narrator too has been a source of that delight, as the distinctions essential to modernism are exploded by postmodernism. Kurt Vonnegut, Paul Auster, and Richard Powers all have written works of fiction that include characters bearing the author's name and some of his characteristics; they deliberately conflate these different versions of themselves. Nabokov is especially cunning in hiding figures of himself within his fictional texts, even to the point of including anagrammatic forms of his name ("Vivian Darkbloom"). In some instances ("First Love," "Mademoiselle O"), Nabokov has published essentially the same text both as fiction and as autobiography, while Bend Sinister (1947) ends with by the protagonist s metafictional intuition that he is returning to the bosom of his creator (see Richardson, "Nabokov's"). Such interpenetrations point to the inherent constructedness of our notions of the author or implied author as well as all notions of the self. They also identify and reaffirm fiction's status as fiction.
The last half of the twentieth century has seen an explosion of experiments with narrators and narration. Although many of these experiments have been tried before, postmodernism has returned to them and explored their capacity for defamiliarizing our perceptions and otherwise responding to our contemporary world. Modern authors have employed first-person plural ("we" narration), third-person plural (where the persons depicted are nearly always multiple and referred to as "they" or "them"), passive voice narration in which the speaker and/or agent remains anonymous and unidentified, and multiple kinds of narration within the same text. All of the forms just mentioned appear in Joseph Conrad's The Nigger of the 'Narcissus' (1899), including the regular juxtaposition of first- and third-person plural accounts, as a position of objectivity does battle with the fact of human subjectivity. Other works start in one form and end up in the other, as what seems to be an omniscient third-person perspective turns out to be the voice of a single individual, as in Iris Murdochs The Philosopher's Pupil (1983). In the contemporary novel, Carlos Fuentes alternates among first-, second-, and third-person narration in The Death of Artemio Cruz (1962), as does Nuruddin Farah in Maps (1986).This kind of multiperson narration forces together opposed narrative perspectives that mimetic authors rigorously keep separate. Another compelling modern narrative stance is
t
i
second-person narration, where the protagonist is referred to in the second person. In Mary McCarthy's "The Genial Host" (1942), we read: "Now you hesitated, weighing the invitation. Sooner or later you would break with him, \ ou knew. But not yet, not while you were still so poor, so loverless, so lonely" (163). Second-person fiction cannot easily be reduced to either homodi-e„etic or heterodiegetic status, but hovers ambiguously between these two 11 mventional positions. Narrative theory needs a more capacious and flexible model of narration in order to circumscribe the full range of modern and 11 ^temporary practices.
The practice of unreliable narration, in which it is apparent that the narrator is deficient in factual knowledge, interpretation, or judgment, has been clearly set forth by Phelan and Rabinowitz. These categories, however, are all based on a mimetic model of narration, that is, of human-like narrators making typically human distortions as they narrate. What happens when an antimimetic author employs this strategy? Here is Saleem critiquing his own reliability: "Why did Saleem need an accident to acquire his powers? Most of the other children [born the same day] didn't." He points out that sometimes "Saleem appears to have known too little, at other times, too much" (530)—this formula is a good description of antimimetic unreliability, by the way. Postmodernism produces still more extreme kinds of unreliability. We have a "fraudulent" narrator when the stated narrative situation is impossible, as when a child displays verbal powers available to only a few adults. A "contradictory" narrator presides over a text that is a mass of contradictions (Robbe-Grillet's Jealousy [1957]), and a "permeable" narrator has a mind that, like Saleem Sinai's in Midnights Children, can be invaded by the private thoughts of others. A "dis-framed" narrator is one whose discourse violates the narrative frame that is supposed to contain it, as when a character narrator claims to have written other books by the author that created him. Each of these possibilities can be seen as radical extensions of the practice of the unreliable narrators of romantic and modernist narratives, but in each case these practices are pushed into the realm of the impossible. The reader is thus shown the fabricated nature of the fictional narrator and, by extension, the artificiality of mimetic limitations that realism imposes on his or her narration. Fictional narrators can do virtually anything, and it is the goal of antimimetic fiction to radically extend practices far beyond what has been expected, imagined, or considered possible.
Antimimetic narrators often include unusual, unnatural, or nonhuman speakers. These include the deranged narrators of Beckett and Nabokov, the mute narrators of Calvino's The Castle of Crossed Destinies (1969), the dead narrator of Beckett's "The Calmative" (1946), a narrating horse in John
2: AUTHORS, NARRATORS, NARRATION
NOTES
Fowless Sweet William (1993), and story-telling machines in Stanislaw Lems The Cyheriad (1967). Postmodern narration can become even more unstable as it presents divergent or disparate textual fragments that seem impossible for a single human individual to have composed. Beckett's "Unnamable" is an especially antihuman narrator: its identity is constantly fluctuating, inconsistent, and self-contradictory. We can conclude that the death of the traditional narrator is a key precondition for the creation of new forms with other, disparate, decentered voices. A simple humanistic framework cannot begin to encompass such acts of narration.
Rushdie plays impressively with the conventions of ordinary narration in Midnight's Children. On the last page of the book, Saleem wishes for grammatical options that exceed the normal three persons: "I have been so-many too-many persons, life unlike syntax allows one more than three" (533). But this limitation is hardly a problem for this narrator. He is, after all, able to read the minds of other individuals born on the same day, a possibility normally reserved for omniscient novelists. His mother, it might be noted, seemed to have similar powers and was said to be able to eavesdrop on her daughters dreams (58). At one point in his narrating, he is so ashamed of his actions that he refuses the first-person pronoun: '"I am glad,' my Padma says, 'I am happy you ran away.' But I insist: not I. He. He, the Buddha. Who ... would remain not-Saleem" (414). At other times, he narrates as if his novel were a film: "Close-up of my grandfathers right hand: nails knuckles fingers all somehow bigger than you'd expect" (30). Throughout the text, Rushdie both evokes and transgresses the possible knowledge of a human narrator and often winds up with extrahuman powers. Once again we see a refusal of conventional parameters, an insistence on creating new, unusual forms, an underscoring of the artificiality of all narrative construction, and an assertion of the power of fictions in human existence.11
MOTES
1 For additional discussion of our views of the implied author, see Peters "'The Absence of Her Voice from that Concord': The Value of the Implied Author" and Jims "The Implied Author, Deficient Narration, and Nonfiction Narrative: Or, What's Off-Kilter in The Year of Magical Thinking and The Diving Bell and the Butterfly?." [JP/PJR]
2. The distinction is not a binary one, since the criteria for determining unreliable narration are of a different order from those for determining restricted narration. Unreliable narration is a function of distance between implied author and narrator, while restricted narration is a limitation on the narrators activity. Thus, restricted narration can be either reliable or unreliable. [JP/PJR]
3. For a fuller discussion of conversation as narration see Jim's "Rhetoric, Ethics, and Narrative Communication; Or from Story and Discourse to Authors, Resources, and Audiences." [JP/PJR]
4. A docent at the Jane Austen Centre in Bath recently showed museum visitors both Cassandra's portrait and the prettied-up Victorian revision of it, explaining that the latter circulates more widely because "the family always thought Cassandra's portrait of Jane didn't look very flattering, or very feminine" (July 7, 2009). That is precisely the quality of Cassandra's drawing that has attracted feminist critics, for reasons well laid out by Margaret Kirkham. [RW]
5. Elsewhere I have developed an extended argument about the relationship between Anne's placement as focal character and the agitated state of her body. See "The Look, the Body, and the Heroine of Persuasion" [RW]
6. Here it should be noted that, for his part, Davidson (3-19) characterized explanations that appeal to reasons as a species of causal explanation. [DH]
7. The situation is more complicated than this brief characterization would suggest. As Tomasello stresses, when one performs a communicative action, one intends for someone to recognize not only what one is trying to communicate but also one's intention to perform that communicative act (94-133). See also H. Clark (129-32) and Grice (217), as well as my contribution to Part Two of this volume. [DH]
8. Abbott {Introduction 100-111) places intentional interpretations on the same footing as symptomatic and adaptive interpretations, which involve reading texts as symptoms of the conditions (social, ideological, or other) out of which they arise or as the starting point for radically transformative retellings, respectively. However, I would argue that intentional readings are more basic or fundamental than the other two types. You can interpret a text as a narrative without reading it symptomatically or adaptively; but you cannot read a story symptomatically or adaptively without having first established that it is a narrative produced for particular reasons—reasons that you then work to background, or bracket, in order to read for symptoms or else engage in transformative adaptations. [DH]
9. In my response in Part Two, I argue that in order to characterize the world-building activities cued by such nested communicative situations, one need not appeal to the multiple reading positions described—and I would suggest reified—by theorists who draw on the narrative communication diagram. [DH]
10. I discuss the concept of "person" in more detail in chapter 5, in connection with issues of character and characterization. As I note there, work by Peter Hobson, P. F. Strawson, and others suggests that part of what it means to be a person is to have a mind as well as a body, that is, a constellation of mental and material predicates. More
2: AUTHORS, NARRATORS, NARRATION
than this, the idea of person entails that mental predicates will be self-ascribable in one's own case and other-ascribable in the case of others. In turn, because narrative texts result from actions performed by that subclass of persons known as authors, it is arguably built into the experience of narrative to engage in (defeasible) ascriptions of mental predicates to story creators—that is, to other-ascribe to authors reasons for their textual performances. [DH]
11. For further reading on these issues, see Hansen, Iversen, Nielsen, and Reitan; Heinze; Nielsen; Makela; and Richardson, Unnatural Voices. [BR]
3
Time, Plot, Progression
James Phelan and Peter j. Rabinowitz
In the introduction we identified attention to narrative progression as one of the core elements of the rhetorical approach, noting that we regard the progression of a narrative—its synthesis of textual and readerly dynamics— as the key means by which an author achieves his or her communicative purposes and the study of progression as a key source of insights into understanding how a narrative creates its effects. As we elaborate on these points and analyze the progression of Huckleberry Finn, we also seek to contribute to the long-standing debate about Twain's treatment of race in the novel. We $ begin with an explanation of why we prefer the term progression to the trail        ditional term plot.
f
PROGRESSION VS. PLOT
IDefinitions of plot range from minimalist ones that make it synonymous with fabula—the chronological sequence of events in a narrative—to maximalist ones that characterize it as the larger principle of organization of a narrative (see, for example, Crane; Peter Brooks; and Ricoeur). But even the maximalist definitions give pride of place among the elements of nar-I        rative to events, their ordering, and their interconnections (causal, acci-
|        dental, analogical, etc.) Our concept of progression arises from a different
I