The primary problem with the notion of the implied author as agent comes down to a confusion between two different kinds of agency--that of creation and that of narration

—an understandable confusion, given that the creator of a narrative also creates its narrators and that the narrator of a text may pose as its creator. Is there then any relevance to the concept of the implied author as a creative principle?

Although Genette’s dismissal of the implied author raises questions that must be dealt with, it still strikes me as rather glib. What, for example, of epistolary novels or novels with two or more narrators, such as Bleak House, The Sound and the Fury or As I Lay Dying? What of novels that include extradiegetic materials, such as John Dos Passos’ USA trilogy with its biographies, "Camera Eye" stream-of-consciousness segments, and "Newsreel" segments? Unless we want to stretch definitions rather severely, there is clearly no one super-narrator who controls such a text, who has collected these letters, combined and juxtaposed these narrations and other elements with each other in the text. Does it then make sense to ascribe these elements and their combination simply to the historical Tobias Smollett, Charles Dickens, William Faulkner, and Dos Passos? Beyond the obvious fact that these people (as far as we know) actually wrote the words of these novels, the structural principle that binds the disparate elements together in each work, that formulates their structural and textual—if not strictly narrative—unity, is something different. A reader’s awareness of connection, coherence and unity among differing elements (including diverse narrations and nonnarrative elements) in a text is not necessarily conditional on attributing them to the name on the title page, a point we will return to in considering film narrative.

The problem is further illustrated by looking at a different aspect of narrative: the question of the range and depth of knowledge that the text exhibits (what Genette refers to as "focalization"). Characters in novels, E.M. Forster wisely observed, are "word-masses" (71). In novels, characters exist only through language. Yet, Forster continues, "[A character] is real when the novelist knows everything about it. He may not choose to tell us all he knows. . . . But he will give us the feeling that though the character has not been explained, it is explicable, and we get from this a reality of a kind we can never get in daily life" (97-98, my emphasis). Forster is wrong about one thing here: It is not the novelist who "knows" everything about the character. We may indeed get the sense that a character is completely explicable, and that the diegetic world of the novel is complete and unified, even if all possible diegetic details are themselves incomplete (as they are, by necessity, in the depiction of virtually every diegesis). A "non-focalized" narrator may feign complete knowledge of a diegetic world and its characters but it can never reveal all that it "knows." On the other hand, what the actual, living, historical novelist "knows" about those characters and that world, outside of the text itself, is quite irrelevant to a reader’s perception of what might be known about that world (even though such knowledge might serve as a supplement to textual knowledge). In fact, writers themselves will sometimes admit their own ignorance of their fictional worlds. Novelists often talk about how a minor character in an early draft "ran away" with the novel and became the book’s major character or how a story developed "on its own" in an unexpected way. Even when a novelist provides a great amount of detail about a relatively small fictional universe, he or she can never say everything. Even with his maps and capsule biographies, what William Faulkner as a person "knew" about Yoknapatawpha County is, again, quite irrelevant. It is enough that the narrators of the stories and novels imply a much wide geography, history, and culture, not to mention the biographical and psychological details of that land’s inhabitants. From these narrations (and Faulkner’s maps and capsule biographies), we infer or reconstruct a model of that "postage-stamp of ground."

Thus, when we infer the completeness and coherence of that diegesis, we also assume an agency that has "intended" that completeness and coherence, whether the historical author had such intentions or not. When a text "knows" more than its characters or even its narrators reveal (as in The Good Soldier), we still know that some agency within the text is prompting us to "fill in the blanks," to "read between the lines." That agency is what might be called the "implied author," or better—as Chatman suggests—"inferred author" of the text. Despite calling this entity an "author," it need not be ascribed to some anthropomorphic entity. Umberto Eco, calling this agent the "model author," describes it as "a narrative strategy, a set of instructions . . . which we have to follow when we decide to act as the model reader" (15). (The "model reader," too, in Eco’s system is a textual strategy, not a metaphysical entity.) The model author is a strategy found in all types of text, according to Eco, "even in the most squalid pornographic novel, to tell us that the descriptions we are given must be a stimulus for our imagination and for our physical reactions" (17).

Even if extratextual sources (say, an author’s notebooks or early drafts) were to suggest different intentions than those we have inferred, they would not necessarily invalidate the "model author" we perceive. We could argue that the historical author did not really understand his or her own works, that the model author we have inferred from the text reveals those intentions more accurately, and that those extratextual sources only reveal the extent of the real author’s blindness or confusion. Much psychoanalytic and ideological criticism in fact proceeds from such assumptions.

Finally, and most relevant to my own point here, print fiction can openly invoke an anthropomorphic implied author through the text’s narrator(s). At the end of Vladimir Nabokov’s Bend Sinister, the hitherto-nameless, faceless nonfocalized narrator takes on the virtual form of what Nabokov himself calls "an anthropomorphic deity, impersonated by me" (xviii). The heterodiegetic narrator ("I") suddenly gets up "from among the chaos of written and re-written pages," distracted by a moth hitting his window screen (216-217). Nabokov, who knew an awful lot about a lot of things, still does not know everything, even about his own books, and is thus required to "impersonate" that implied deity. When the "author" posts a warning at the beginning of Huckleberry Finn that no motive, moral or plot is to be found in his book, and that warning is immediately followed by the narrator Huckleberry Finn claiming that he is the author of this narrative and that "Mr. Mark Twain" had told "stretchers" about him in Tom Sawyer, we are caught in a game of logical paradox that cannot be ascribed to either "Twain" or Huck alone. "Twain" and Huck challenge the mutual reliability of the putative author (Twain) and the "author"-narrator (Huck) and thus provoke the reader not only to see the absurd humor of the paradox but also infer that perhaps some motive, moral, or plot might in fact be found here. If we want to name a source for this paradox, we surely cannot call it "Samuel Clemens," but we might perhaps call it the "implied (or inferred) author."

return to top