Consider the following passage from Sherwood Anderson's "Death in the Woods":

There was such an old woman who used to come into town past our house one summer and fall when I was a young boy and was sick with what was called inflammatory rheumatism. She went home later carrying a heavy pack on her back. Two or three large gaunt-looking dogs followed at her heels. (1211)

Using first-person, the narrator depicts a scene as he remembers it; yet, the rest of the story contains scenes at which he could not have been present:

She managed to get out of the house without her employer seeing, but when she was getting into the buggy he showed up. It was almost dark, and he just popped up suddenly at the horse's head. He grabbed the horse by the bridle and Jake [the employer's son] got out his buggy whip.

They had it out alright! The German was a tough one. Maybe he didn't care whether his wife knew or not, Jake hit him over the face and shoulders with the buggy whip, but the horse got to acting up and he had to get out. (1212)

Recognizing that these few terms describing point of view fail to adequately represent the complex and varied relationships that occur in narratives, critics in narratology—a special field of literary criticism—argue about the metaphors and terms we use to describe the relationship between narrators and characters. Such is the case in Harry Shaw's essay "Loose Narrators" in the May 1995 issue of Narrative where he exposes problems in the "crisp"¹ distinction between "story space" and "discourse space" in Seymour Chatman's Coming to Terms.

My purpose is to describe more pointedly the relationships between narrators and characters by exposing problems in the models of both Shaw and Chatman and by defending and further developing an alternative metaphor, known as focalization ...

Chatman uses the space metaphor to distinguish between the time and place of a narrator telling the story and the time and place of the characters experiencing it. Shaw argues that particularly "vibrant" (99) narratives and narrators "wishing to enter the historical world of the characters" (104) seem to elude the metaphor of a "membrane" (Chatman 144) separating narrators and scenes. Chatman defends his position in the October 1995 issue of Narrative, which includes a subsequent response by Shaw. While Chatman holds to his distinction, emphasizing that it allows us to describe more pointedly the "genuine exceptions" (305), Shaw claims that he did not intend to imply that the distinction should be abolished or that it is not important (309), but rather that it fails to adequately describe all narratives. My purpose is to describe more pointedly the relationships between narrators and characters by exposing problems in the models of both Shaw and Chatman and by defending and further developing an alternative metaphor, known as focalization, which has been introduced by Gérard Genette and Mieke Bal among others.

I specifically wish to address two of Shaw's arguments: one involves the energy and vividness of a scene, and the other involves the "historicity" of a narrator. Shaw suggests that at times a narrator seems to enter story space as if she is "imitating the role of someone who has happened upon a scene, is struck by it, and possesses the rare facility of being able to capture it in a spontaneous, instantaneous image that crystallizes the experience . . . it embodies" ("Loose" 99). This claim does not seriously challenge Chatman's distinction between the two spaces. I may imitate the role of the boy in the bubble, but I am never actually inside the bubble, and the fact that I am imitating, as Shaw himself has argued ("Thin Description" 309), means that I am fully aware that I am not inside the bubble. Perhaps Shaw simply points out that some narrators seem to appear as though they invade story space. If so, who would argue otherwise? Is that not, after all, the great thing about fiction? Its illusiveness?

In addition, Shaw does not suggest that the characters in the scene imitate the role of people who suddenly recognize someone who has happened upon them, or is imitating the role of someone who has happened upon them. In order for Shaw's description of the appearance of a story-space-discourse-space invasion to be complete, people on both sides of the "membrane" ought to have equal access to the recognition of--or the pretense of recognition of--the invasion. In other words, if discourse is not separated by a membrane of sorts so that the narrator can exist in the same realm with the characters in the story, the characters, then, should be able to see the narrator, to interact with him. In Shaw's scenario, the narrator appears privileged, as if one of those one-way mirrors psychologists use to observe their subjects prevents the characters from noticing the narrator. In other words, there seems to be no mutual sharing, or an imitation of a mutual sharing, of the same space. For example, in the muffin scene from Dickens's Dombey and Son—much discussed in the Chatman-Shaw debate—Son remains "tucked up warm" in his little basket while the narrator makes his muffin analogy, and Dombey does not wonder at the stranger who has come to contemplate them.

Therefore, Shaw does not convince me that narrators metaphorically appear to share story space with characters because he does not establish that the recognition of the invasion by narrators is reciprocated ...

This absence of reciprocation also challenges Shaw's claim regarding a narrator's "historicity." For example, when Shaw discusses passages from George Eliot's Scenes of Clerical Life, he asks if the narrator's "emotional engagement" implies that "she is edging toward story space" ("Loose" 106). Again, edging toward and invading are not the same thing, but suppose Shaw means that story space is invaded by the narrator. Even then, Tina, who the narrator describes as carrying overwhelming burdens, and who then would be sharing the space with the narrator, seems unaware that a sympathetic presence shares them:

In that summer, we know, the great nation of France was agitated by conflicting thoughts and passions, which were but the beginning of sorrows. And in our Caterina's little breast, too, there were terrible struggles. The poor bird was beginning to flutter and vainly dash its soft

breast against the hard iron bars of the inevitable, and we see too plainly the danger, if that anguish should go on heightening instead of being allayed, that the palpitating heart may be fatally bruised. (147)

As in the muffin scene, this passage quoted by Shaw in "Loose Narrators" illustrates a narrator drawing an analogy, but, as in the muffin scene, Tina takes no notice of another person there. The narrator may be emotionally engaged in Tina's historical element, but she does not share Tina's story space. In fact, Tina's loneliness is emphasized by the absence, in this passage, of a sympathetic presence.

Of course, the narrator could invade Tina's space without her awareness if characters lack the independent agency of real people. But if we consciously describe the phenomenon that Shaw means to pinpoint in "Loose Narrators" with this kind of relationship between a powerful narrator and a fictive character, then the illusion of fiction is broken, and the invasion of story space draws attention to itself. I do not believe that Shaw means to suggest that Tina's narrator and the narrator of Fowles's The French Lieutenant's Woman or of Julio Cortázar's "Continuity of Parks" have this artificiality in common. Therefore, Shaw does not convince me that narrators metaphorically appear to share story space with characters because he does not establish that the recognition of the invasion by narrators is reciprocated by characters or explain why such a recognition would not occur between people sharing the same space. Even a fly on the wall is usually eventually noticed by those in the same space with it. But Shaw does not want a fly or any other dehumanized teller; he wants a human. And a human on the wall would definitely attract a character's attention.

Shaw seems concerned with energy and history, but Chatman's distinction is not about either of those. It should be permissible for Chatman to define a particular element of fiction without having to explain how scenes are more or less energetic or vivid, or how narrators more or less long for the historical setting of their scenes. But Shaw's agenda appears to involve a challenge to more than Chatman's space distinction. He seems adamantly against a precise terminology for describing narrative. He appears concerned that the majestic illusiveness of fiction will fade in the "crisp" dissection. He says that he "prefer[s] to view analytical terms as provisional and shifting, as a set of approximations" ("Thin" 312). A homogeneous vocabulary for the study of narrative could be nothing but entirely arguable and debatable in its application to specific texts. But we need a vocabulary before we can approximate, although it may include synonymous terms such as "story," "plot," and "fabula." We may continue to argue about the terms, but we should still have them, and they should be as accurate in their meaning as possible.

And could we not imagine in such a case that the narrator, on the opposite side of the membrane from the scene, is capable of--no, must be capable of--witnessing it if he is to accurately "report" it?

It is this hope for precision that brings me to Chatman. Obviously my purpose is not to complain about his story-space-discourse-space distinction. Like Shaw, however, I see problems with the metaphor, not because I think it fails to recognize particularly energetic scenes or historical connections between narrators and characters, and not because I think the division renders a kind of cold objectivity in the discussion of narrative, but because space metaphors, like time metaphors, fail to adequately describe the relationships between narrators and their stories.

Although Chatman, in Coming to Terms, more specifically discusses the spatial boundaries between characters and narrators, he refers to the temporal ones briefly when he describes the narrator in Dombey and Son. Chatman explains that the narrator "resides in an order of time and place different from that occupied by the characters; his is a different 'here-and-now.' And that's true for every narrator, no matter how minimal his/her/its distance from the 'here-and-now' of the story" (142). Chatman argues that, because of these spatio-temporal differences, the narrator is not to be imagined as an "observer" of the scene (142). But how temporally separate are story and discourse during simultaneous narration in which the narrator uses present tense to describe the events of a story? Consider, for example, Chekhov's "Children" from The Cook's Wedding and Other Stories:

Papa and Mama and Aunt Nadya are not at home. They have gone to a christening party at the house of that old officer who rides on a little grey horse. While waiting for them to come home, Grisha, Anya, Alyosha, Sonya, and the cook's son, Andrey, are sitting at the table in the dining-room, playing at loto. . . . The table, lighted by a hanging lamp, is dotted with numbers, nutshells, scraps of paper, and little bits of glass. (25)

Although a membrane of some sort still separates the two diegetic levels, do they not coexist in the same time? And could we not imagine in such a case that the narrator, on the opposite side of the membrane from the scene, is capable of--no, must be capable of--witnessing it if he is to accurately "report" it?²

It may seem safe to say, then, that story and discourse may share the same temporal plane but not the same spatial one. But one may imagine another Dickens text, A Christmas Carol, in which the ghosts of Christmas past and future and Ebenezer Scrooge share the same space as characters who do not see them while existing on different temporal planes. Imagine a similar relationship between narrators and characters. I argued earlier against Shaw's notion of "loose narrators" that narrators and characters must mutually recognize an invasion of story space, but what if they share separate temporal planes while existing in the same space? Both past tense and future tense narration can be viewed through this metaphor. Imagine a narrator at the scene he or she describes but existing in a different time. Character recognition of the narrator's presence would no longer be a factor. Has story space been invaded?

Perhaps a less confusing metaphor for distinguishing between discourse and story is Genette's concept of the three diegetic levels: extradiegetic (the level of the narrator's telling), diegetic (the level of the characters and their thoughts and actions), and metadiegetic (a story within a story, as when a diegetic character tells a story). The word "level" is not necessarily a temporal or a spatial metaphor, although one can progress to later levels in a sequential task and take elevators to higher and lower levels in a building. With these three levels, narrating remains distinct from narrative in that the first occurs on the extradiegetic level while the second is the sum of all three. It is less ambiguous (though perhaps more cumbersome) to say that an extradiegetic narrator does not function on the diegetic level than it is to say that it does not penetrate story space. Exceptions to this general rule of narrative levels are called metalepses, which occur when narrators or characters on one diegetic level appear on another, as in Julio Cortázar's "Continuity of Parks" when the characters in a story the protagonist is reading appear behind his chair (Narrative Discourse 234-35). In "How Loose?" Chatman describes the metalepsis in John Fowles's The French Lieutenant's Woman when the narrator recognizes himself in the story and then suddenly becomes a homodiegetic (first-person) narrator (304-305).

To further illustrate the subsequent confusion that results in using the space metaphor, let us consider Chatman's replacement for "point of view." In order to distinguish between story space and discourse space, Chatman introduces the terms "slant" and "filter" as more precise replacements for the "points of view" of narrators and characters. A narrator residing in discourse space provides the "slant" or "angle" to the story, while some of his or her information may be "filtered" through one or more characters existing within the story space (143). If "filter" means a porous mechanism that allows some things and not others to pass through, like a coffee filter or an air filter, then I imagine a discourse which is "filtered" through a character as one that is somehow incomplete, not wholly the narrator's discourse. Chatman, I believe, with the term "filter," empowers characters with the ability to silence some of the narrator's discourse. This is not a bad thing. However, the threatening implication to Chatman's distinction is that if characters are "filtering," then the narrating is passing through the "story space" like water passes through a coffee filter. This filtering confuses the when and where of the story, in my opinion, more than the "point of view" metaphor Chatman wishes to right.

External and internal focalization describe two points of view a narrator may take.

Chatman argues in Coming to Terms that the phrase "point of view" has multiple meanings such as "perspective," "stance," and "interest" (140-41). He maintains that the mental functions of narrators and characters are different, and so one term, such as focalization, should not be used to describe both (141). But I believe a modification of the way we have been using the term focalization would answer Chatman's concerns.

Genette, like Chatman, makes the distinction between point of view and voice, describing the "confusion between the question who is the character whose point of view orients the narrative perspective? and the very different question who is the narrator?" (Narrative Discourse 186). He differentiates between, for example, the narrator of William Golding's Pincher Martin and Chris, the character from whose point of view we see most of the story. However, another distinction becomes obscured in Genette's discussion: the difference between point of view and focal point, or the difference between seeing from the perspective of a character and having that character be the focal point from a perspective outside of the character.

This distinction becomes confused because Genette uses the concept of focalization to describe the focal point of a narrative and not the point of view. He defines internal focalization as the instance when "the focus [or focal point] coincides with a character, who then becomes the fictive 'subject' of all the perceptions, including those that concern himself as object" (Revisited 74). He describes external focalization as the circumstance when "the focus [or focal point] is situated at a point in the diegetic universe chosen by the narrator, outside of every character, which means that all possibility of information about anyone's thoughts is excluded" (75). He says in Fiction and Diction that external focalization "consists in abstaining from any intrusion into the characters' subjectivity, reporting only their acts and gestures as seen from the outside" (66).

While Genette's concept may answer the question what is the focus of the narrative?, it does not answer the question he originally suggests needs answering in Narrative Discourse: who sees? (186). External and internal focalization describe two points of view a narrator may take. Two other possibilities exist: 1) a narrator may see from the external point of view of a character, and 2) a narrator may see from the internal point of view of a character.

Focal point differs from point of view in that the narrating includes references to that character, which keep the perspective outside of the character. For example, if a narrator were to say: "She saw the wild horse," the narrator sees the character seeing the wild horse and that character remains an external focal point. If the narrator were to say: "The wild horse came into view," then we would be seeing the wild horse from the point of view of the character so that the wild horse—and not the character—is the external focal point. So, in the first case, the focal point is the character—her physical act of seeing the wild horse—through the single vision (or angle or slant) of the narrator, and the narrator, as Meike Bal has clarified, is the focalizer (106). In the second case, the focal point is the character's view of the world—from her physical standpoint—through the double vision (or perspective) of the narrator and character, so the character becomes a focalizer and her view of the world, specifically the wild horse, is the focalized (Bal 106).

The same distinction can be made between "She thought, 'What a beautiful wild horse'" and the character's untagged thought, "What a beautiful wild horse." In the first instance, the character's thoughts present us with an internal focal point—they remain objectified by the tag. In the second instance, the thoughts become an internal point of view—no longer objectified. The stream of consciousness puts us completely into the shoes of the character.

To limit the concept of focalization to describing internal and external focal points is to fail to describe who sees or whose point of view orients the narrative perspective.

Mieke Bal differentiates between the focalizer and the focalized, or the subject and the object, or the seer and the seen, and she explains that both narrators and characters can be focalizers. She signifies character focalizers with "CF" and narrator focalizers with "EF" (because they are extradiegetic focalizers). Her formula can track the shifts in focalization from sentence to sentence or from phrase to phrase. With Bal's formula, a character can change from being a focal point (or the focalized) to being a focalizer within a few words:

She felt the clammy soft hand of the youth.

Although the referent "she" focalizes the character, the rest of the sentence appears to be information that comes from the character's standpoint; thus, the focal point shifts from the character "she" (focalized by the extradiegetic narrator) to the youth's hand (focalized by the "she").

So how do we apply focalization in a broad way so that we can avoid diagramming every sentence of a novel?

First, we distinguish between types of focalization. A narrator who tells a story without the points of view of other characters presents a nonfocalized story (this occurs when a character or characters remain focal points and are never focalizers; the narrative is only focalized extradiegetically). Second, when a narrator focalizes either externally or internally through a character, even if that narrator frequently refers to that character, we should call this a focalized story. Third, borrowing from and modifying Genette's terms, we can distinguish between fixed and variable focalization (Narrative Discourse 189-90). Fixed occurs when a single diegetic character functions as a focalizer, such as Stephen in James Joyce's A Portrait of the Artist as a Young Man. Here the use of fixed focalization enhances the theme of the solitary individual alienated from the rest of humanity. Variable occurs when two or more diegetic characters alternately function as focalizers, such as in F. Scott Fitzgerald's Tender Is the Night, where the perspective alternates between that of Rosemary, Nicole, and Dick among others to create a decentered text about the decentered personality and the crumbling national identity in America. Variable focalization also occurs in Margaret Drabble's The Gates of Ivory,³ in which the sense of a fragmented and arbitrary world is undercut by a prevailing and perhaps fatalistic sense of structure. These categories—non, fixed, and variable focalization—can be applied to an entire novel in a general way so that the label represents a novel's predominant use of point of view.

It is through focalization that texts create the illusion that narrators share the same space as the characters. Narrators appear to be so close, in fact, that they seem to inhabit the bodies and minds of characters.

In other words, if we understand, for example, that The Gates of Ivory uses variable focalization, then we understand, even if there may be some sections which appear fixed or nonfocalized, that the narrator has greater privilege than those who predominantly focalize through one character. The narrator employing variable focalization has greater privilege because he or she can present the reader with various diegetic perspectives. We should accept focalization as a privilege because, except for the homodiegetic (first-person) narrator who focalizes through a younger version of himself or herself, it cannot occur—unless it is conjecture and speculation—outside of narrative. The most privileged narrator, then, is that of the variably focalized text while the least privileged is that of the nonfocalized text. An example of the latter occurs in Herman Melville's Billy Budd, Sailor where a heterodiegetic (third person) narrator speculates from a considerable temporal and spatial distance the motives of Billy, Claggart, and Captain Vere.

Narratives told by heterodiegetic (third person) narrators may employ variable, fixed, or non focalization, but texts told by homodiegetic (first person) narrators cannot employ variable focalization. And the fixed focalizer employed by homodiegetic narrators is limited to that version of himself or herself that exists in the story. Exceptions may occur with artificial narrators who foreground the synthetic component of their stories.

It is through focalization that texts create the illusion that narrators share the same space as the characters. Narrators appear to be so close, in fact, that they seem to inhabit the bodies and minds of characters. This is why Genette's diegetic levels describe better than spatial or temporal metaphors the relationships between narrators and scenes.

So, basically in accordance with the limitations Chatman seems to have wanted to put on narrators, extradiegetic narrators generally do not exist on or in other diegetic levels: the "membrane" remains impenetrable; however, in accordance to what Shaw seems to have wanted to emphasize, the focalization employed by extradiegetic narrators provides the immediacy of perception that makes them appear as if they do.

¹ Shaw actually uses the word "crisp" in his later essay "Thin Description" published in Narrative, October 1995, page 308, in which he defends the earlier essay.

² Chatman prefers "report" to "contemplate" when describing the narrator's particular function in the muffin example from Dickens's novel (Coming to Terms 141-42), and Shaw questions this preference in his "Loose Narrators." It seems to me that narrators can be capable of contemplation, even in posterior narration; for, after all, narrators can imagine scenes before them as they tell them, even without sharing the same "here-and-now," and, in their act of telling, may contemplate certain aspects of scenes. We contemplate episodes in our past all the time, do we not?--even in, or especially in, the process of recounting those personal episodes to other people. What I mean to say is that the act of contemplation does not require an immediate observation of the subject being contemplated. One may contemplate the future of a democratic nation, the progress of AIDS research over the next few decades, a movie one saw two months ago.

³ I suggest that we discard "multiple" focalization, which Genette distinguishes from "variable," because there is almost always a chronological overlap in variable perspectives, and the difference may not be significant in a broad application to a novel.