The Encyclopedia of the Novel - Peter Melville Logan 2014


Paul McCormick

Narrator refers to the mediating agent through whom an author presents a narrative. To the question, who tells?, the answer is always “the narrator.” (However, there is a minority position that argues for versions of the “no-narrator theory,” explained later.) But importantly, it is not always the answer to the other key question of narration in a novel, through whose perception do we understand the story? That is a question of point of view or focalization, for the narrator may tell the story not through his or her own perspective, but rather through those of characters in the story's world.

Authors assign narrators specific features in order to achieve specific effects, and much research about narrators entails distinctions among their possible features. In fact, a more specialized definition of a narrator is: a collection of various features (traits, beliefs, ethics, linguistic habits, and ultimately functions) assigned by an author to a designated storyteller. Distinguishing among different types of narrators allows readers to better understand the selection of features from which authors can choose and why they select and combine certain features. In general, two critical concepts have proven particularly useful in conceptualizing the possible relations among those features and effects: the Proteus Principle and the concept of narrative situations. Meir Sternberg's Proteus Principle states that “there are no package deals in narration” because there are “many-to-many correspondences between linguistic form and representational function” (1982, 112). With respect to narrators, the Proteus Principle indicates that any particular narrative feature may lead to a wide range of narrative effects because the effects depend not just on that feature but also on many other elements of narrative. In a way, the Proteus Principle helps to qualify and balance the concepts of narrators and of narrative situations as previously developed in classical studies by Gérard Genette and Franz Stanzel. For Stanzel, a narrative situation conceptualizes narrators as bundles or arrangements of different features relating to their identity, point of view, and degree of intrusion. While different features can lead to different effects, understanding how different features are often bundled together allows readers to compare the similarities and differences among different types of narrators and offers a point of reference for generalizing about the sort of effects authors have historically achieved with different combinations of narrator features.

Accordingly, this entry begins by describing two fundamental ways of distinguishing among narrators: identifying a given narrator's participation in story and the level of her narration in relation to the primary action-level (Genette, 227—62). Then, the article uses three sections to discuss Stanzel's three narrative situations: authorial, figural, first-person—and some larger issues related to each.

Narrator Participation and Narrative Levels

A fundamental distinction of novelistic technique is whether a given narrator is participatory (and physically present) or non-participatory (and physically absent) in the story she is telling (see NARRATIVE TECHNIQUE). In the past and less frequently today, this distinction was often roughly made by both authors and critics who relied on a grammatical opposition between “third-person” and “first-person” narrators. However, as Genette points out, this grammatically based taxonomy is too imprecise because any narrator can use the first-person and almost all use the third. Genette suggests that a better way of making the appropriate distinction is to distinguish between narrators who are able to participate in the narrated action (homodiegetic) and those who are not (heterodiegetic). In addition to participation, Genette identifies narrative levels as another key variable influencing a narrator's telling. Here, I prefer the term external narrator to replace heterodiegetic narrator (third-person) and character narrator to replace homodiegetic narrator (first-person).

The distinction between external and character narrators is essential because it is tied to their respective epistemological privileges. The storyworld non-participation of external narrators can correlate to a privileged and even unworldly knowledge of characters and events; e.g., some external narrators have full and unmediated access to the interior mental and emotional states of several characters. In contrast, the realistic conventions of character narration usually demand that these narrators restrict their reports to what they witnessed or can retrospectively infer from their experiences in the storyworld. Character narrators can be very knowledgeable indeed, even with respect to the inner lives of other characters. However, character narrators' special knowledge of the inner workings of other characters must be justified (i.e., motivated, naturalized) or readers may suspect their claims of knowledge. In contrast, it is a literary convention that external narrators may have complete and reliable access to the inner lives of characters without explanation (see DISCOURSE). If an external narrator quotes a character's thought, readers typically take the quotation as wholly accurate.

The concept of narrative level places acts of narrating (and thus individual narrators) and narrated stories in relation to the entire narrative of which they are parts. There may be many narrating acts and many narrated stories in one novel, and consequently many narrative levels and narrators, which narrative theorists have proposed various terminologies to describe and analyze. Here I draw primarily upon Genette's model to outline a procedural approach for placing narrators on narrative levels. The first step is to determine whether a narrator is an external narrator or a character narrator. The second step is to identify the primary action-level of a novel, often called its diegesis in reference to Genettian vocabulary. For example, Genette uses the term extradiegetic to signal a narrator once-removed from this primary action-level, and intradiegetic to signal a narrator telling a story on that primary action-level. However, I prefer to speak in terms of remove from the primary action-level while retaining Genette's concepts. For example in Gustave Flaubert's Madame Bovary (1857), Rodolphe's seduction of Emma Bovary is part of the primary action-level (pt. II, chap. 9); this narrative level should be distinguished from the one occupied by the external narrator; it should also be distinguished from the world and actions described in Lucie de Lammermoor, an opera which Emma attends (pt. II, chap. 15). The third step is to ask whether the narrator is narrating the main level of action at one remove (a narrator at one-remove) or if the narrating act occurs at the same level as the primary action-level (a narrator at zero-remove). In other words, at how many removes is a particular narrator from the novel's primary action-line?

Once these first three steps are completed, readers can execute the final step of identifying what level a particular external or character narrator occupies. For example, George Eliot's external narrator in Middlemarch (1871—72) operates at a single remove from the primary action-line (external narrator at one-remove). While Joseph Conrad's Marlow functions as a character-narrator in Heart of Darkness (1902), he also narrates at one-remove because he retrospectively narrates the novel's primary action-line (character narrator at one-remove). Both external and character narrators can also narrate at zero-remove. For example, in James Joyce's short story, “The Two Gallants” (1914), Corley narrates on the same narrative level as the primary action-line when he tells Lenehan how he first seduced the maid he will meet later, so Lenehan is a character narrator at zero-remove from the primary action-level (but one-remove from his story about the maid). Compare this to the “Hades” episode of Joyce's Ulysses (1922) when Martin Cunningham tells Mr. Power that Rudolph Virag poisoned himself (6:529). Martin was a non-participant in the story of Virag's suicide, but it occurs on the same narrative level as his current ride to the cemetery. So in this instance, Martin Cunningham functions as an external narrator at zero-remove. These four combinations of narrative features are Genette's version of narrative situations because they represent four common combinations of types of narrators and narrative levels. However, Genette notes that other options exist in novels with multiple narrative levels. For example, narrators can be several times removed from the primary action-line (e.g, narrators at twice-remove). In addition, several narrators can exist at the same narrative level, as in the first three narrators of William Faulkner's 1929 The Sound and the Fury (serial narrators). In each case, however, the primary action-line (the diegesis) is the baseline from which all distinctions regarding narrative level are made.

External Narrators and the Authorial Narrative Situation

The negative correlation between story participation and story knowledge is strong enough that the two most famous studies of narrators differ on what primarily defines an external narrator: Are narrators of novels like William Makepeace Thackeray's Vanity Fair (1848), Anthony Trollope's Barchester Towers (1857), and Henry Fielding's Tom Jones (1749) defined by their non-participation in the story or by their unworldly knowledge (omniscience)? Generally, as in this entry, the distinction between participation and non-participation is held to distinguish external narrators and character narrators, respectively. But when Franz Stanzel offers his classic definition of the authorial narrative situation, he has good reason to discuss a prototypical external narrator (his “authorial” narrator) as one whose unworldly knowledge or omniscience is her primary trait. In his model, the opposite of an authorial narrator is not a character narrator, but a limited point of view. Certain external narrators like those of Vanity Fair and Tom Jones do seem to flaunt their omniscience to the point where it becomes their dominant characteristic, and controlling the knowledge of both characters and readers is crucial to authors' narrative techniques. However as Dorrit Cohn (1978) notes, unworldly knowledge means that the narrator exists out of the world, that in some sense the unworldly perspective of Stanzel's authorial narrator means that she is also Genette's non-participatory narrator. Still, there are gradations of omniscience among external narrators who are primarily defined by their non-participation in the storyworld. While non-participation/participation is key to discerning between external narrators and character narrators, the difference between Genette and Stanzel reminds us that non-participation is often bundled with privileged knowledge, to varying degrees.

Traditionally called “omniscient narrators” in Anglo-American literary criticism are the external, once-removed narrators like those of Vanity Fair, Middlemarch or Leo Tolstoy's Voyná I mir (1865—69, War and Peace) that offer “inside views” of many characters in the storyworld, often commenting on the narrative world and reporting not just characters' actions, speech, and writing, but also their emotions and cognition. As Stanzel observed, these narrators were particularly popular in nineteenth-century Euro-American novels and less popular in the twentieth century. However, Stanzel prefers the term “authorial narrator” to refer to such narrators, and the terminological value of “omniscient narration” has been recently contested. Detractors of the term consider it a sloppy analogy with untenable theological freight because we do not know the characteristics of any deity to make the comparison (e.g., Culler); however, its supporters note that many authors have made the same analogy of “godlike” powers and that some external narrators do exercise unusual and even divine knowledge as mediators of the narrative world (e.g., Olson). Salman Rushdie's novel The Satanic Verses (1988) offers a good example of a contemporary author having some fun with the concept of omniscience when his external narrator says, “I know the truth, obviously. I watched the whole thing. As to omnipresence and -potence, I'm making no claims at present, but I can manage this much, I hope. Chamcha willed it and Farishta did what was willed. Which was the miracle worker? Of what type—angelic, satanic—was Farishta's song? Who am I? Let's put it this way: who has the best tunes?” (pt. 1, sec. 1). Although sometimes, as in the case of Satanic Verses, the connection between some narrators and omniscience deserves exploration, most external narrators offer inside views of only selected characters.

Stanzel's term authorial narration also suggests how the greater epistemological privilege enjoyed by external narrators and removed narrators (once-remove or more) can conventionally signal closer proximity between the implied judgments, norms, and ethics of the author and those of the narrator. But this also helps to explain why twentieth-century authors tend to use external narration less frequently than nineteenth-century authors. In the middle of the twentieth century, Percy Lubbock used the terms “showing” versus “telling” to discuss the same distinction as it pertains to the novel. For Lubbock, the journals of Henry James revealed a prescriptive difference between the two: showing is always preferable to a narrator telling. What Lubbock meant was that a story should be presented as if unmediated by the presence and opinion of a narrator, that a dramatic style of presentation was best. However, Wayne Booth made two influential observations in reply, first, that an omniscient narrator who uses intrusive commentary to comment upon the story is often just as appropriate and artistic for a particular story, and second, that strictly speaking, showing in the novel genre is impossible anyways because some agent must mediate or narrate the action. Today, the distinction between direct and indirect speech and thought representation is not prescriptive but descriptive and often analytically so: identifying whether speech and thought is represented directly or indirectly can often provide important information about the narrator, including the specific relations between the narrator and a given character.

Narrators represent and communicate ethics, history, and politics as well as epistemology when authors select and combine their features. When Mikhail Bakhtin's influential scholarship was first widely received in the U.S. in the 1980s, it became clear that a specialized formal study of narrators could be strengthened by studying the historical and ideological inflections of narrators and their discourses (see IDEOLOGY). As Brian McHale puts it, “Of course, it is precisely his insistence on historicizing language, on restoring it to its place in a historically contingent social realm, that has made Bakhtin so congenial to so many varieties of historicist and contextualist theory in our own time” (63). For example, feminist narratologists have made significant contributions to “the study of narrative structures and strategies in the context of cultural constructions of gender” (Warhol, 21) with a Bakhtinian-inflected dual interest in history and form. Many of these studies have broad implications for the study of narrators. For example, Susan Lanser (1992) has shown how women writers can use the authority conventionally granted to external narrators to establish their discursive authority but also to question the origins of that authority. Robyn Warhol has made an influential distinction between external distancing narrators who discourage the actual reader from identifying with the narratee (the textual recipient of the narrator's telling), and external engaging narrators who encourage actual readers to identify with the narratee. Building upon the foundational work of D. A. Miller and Gerald Prince, respectively, Warhol has recently detailed how classifying and attending to what narrators do not narrate, what she calls “the unnarrated,” often reveals much about authorial purpose, social norms, and genre identification. A wide variety of historical approaches, including those of feminist narratology, have helped to clarify the implications of various narrator features for actual authors and readers.

External narrators can be dramatized to different degrees, and often they do not self-identify their gender, race, ethnicity, sexuality, etc. Ungendered narrators pose an additional practical problem for literary criticism—how should one refer to the narrator if she or he is left ungendered and unnamed, as in Austen's Emma (1815)? As a result, many scholars follow what has become known as Lanser's rule: In the absence of any text-internal clues to the narrator's sex, use the pronoun appropriate to the author's sex. Assume that the narrator is male if the author is male, and that the narrator is female if the author is female (Lanser, 1981, 166—8). This rule is not without its complications, for it adds personal qualities to the narrator that the author apparently did not specify, and disambiguates what the author may have left purposely ambiguous. All the same, it is sometimes awkward to discuss a hypothetical narrator, which is why I alternate between he and she in this article. Lanser's rule makes for easier practical reference and also sets a standard that can be challenged in appropriate cases.

However, the existence of such a rule evokes more significant questions, especially for external narrators. For example, should narrators be interpreted anthropomorphically when there is little textual support for such an interpretive decision? In other words, should one assume that external narrators are always somehow human and attribute to them full human qualities? And why always assume the presence of a narrator (human-like or not) instead of attributing the narration directly to the author (implied or otherwise)? The answers to these questions can depend upon the particular narrative in question: e.g., on the degree of the narrator's consciousness of their narration, whether the narrator offers commentary and judgment, and whether the narrator's voice is distinctive. Toward one end of the continuum one could place external narrators yet personal narrators like those of Vanity Fair or Tom Jones, and on the other end, some of the more impersonal external narrators of Ernest Hemingway's “The Killers” (1927), John Dos Passos's The Big Money (1936), or Alain Robbe-Grillet's La Jalousie (1957, Jealousy).

But answers also depend upon readers' assumptions about narrators. The claim that there is always a narrator in every story largely derives from the assumption that literary narration is a kind of speech or communication act, in which someone must necessarily speak to someone else. That is why many communication models are symmetrical, with an implied author speaking to an implied reader, a narrator speaking to a narratee, and the real author speaking to the real reader, as in the influential communicative model developed by Seymour Chatman (see diagram in NARRATION). Conversely, the claim that there need not be a narrator—called the no-narrator theory—often derives from linguistic analyses in which acts of speech and thought are traced to certain grammatical agents—all of which must exist in the narration's syntax because expressivity is located in grammar (e.g., Banfield). From this perspective, the concept of voices is subordinated to deictic centers, linguistic centers of consciousness whose use of directional and temporal words like “here” and “then” spatially and temporally situate them in the storyworld (see SPACE, TIME). Some no-narrator approaches argue that the narrator is not always an inherent element of narration, while Richard Walsh argues more radically against any necessary qualitative distinction between narrators and characters: “The narrator is always either a character who narrates or the author” (505). postmodern and experimental texts often seem to delight in raising theoretical as well as hermeneutic questions about a narrator's humanity or gender, and recent studies of “unnatural” narrators have brought the possibility of non-anthropomorphic narrators to the fore (e.g., Richardson; Alber). In general, it seems likely that individual authors differ on whether or not their narrators are always anthropomorphic beings, just as readers and theorists do.

Reflectors and the Figural Narrative Situation

The type of narration in which the narrator seems most withdrawn, covert, or absent is often reflector narration, or what Stanzel calls the figural narrative situation. Narratives using this mode of narration can appear to have no narrator at all because the story is told through the perspective of a single character without his or her knowledge. Examples of reflector narration include Joyce's The Portrait of the Artist as a Young Man (1916) and James's The Ambassadors (1903). At first, such novels seem to have neither an external nor a character narrator because the narration offers the ostensibly unmediated thoughts of only one character, but those thoughts are presented in the third person. As Käte Hamburger once noted, it is only in literature that the I-originarity of another's self can be presented in the third person as if from their very own perspective. And in chap. 1 of Joyce's Portrait, for example, we see this I-originarity in the third person without the intrusive presence of an external narrator when we read sentences like, “He had to undress and then kneel and say his own prayers and be in bed before the gas was lowered so that he might not go to hell when he died” (30). But although we thus gain unworldly access to Stephen's thoughts, the third-person syntax reveals that it is not Stephen who tells the story.

But while for some critics this novel may have no narrator, most would say that novels like Portrait are narrated by external narrators but reflected through the consciousness of a particular character. In other words, the “voice” is that of an external narrator who is looking through Stephen's “vision.” Reflector narration is an important subset of external narration, but it is a subset: the narrator is an external, covert narrator who is merely choosing to perceive the world as reflected through a character's consciousness (see NARRATIVE PERSPECTIVE). For authors, the advantages of reflector narration are several. First, readers may be more willing to identify with a character who is not consciously crafting his identity through the narrative. Whatever Stephen's faults, we know he is not performing for the authorial audience. Second, the author may restrict her ideal reader's knowledge more naturally, i.e., with less sense that the author is tricking her. This explains why Stanzel locates the figural narrative situation next to limited point of view. For example, because we see through Strether's consciousness, we are more apt to learn about the affair between Chad Newsome and Mme. Vionnet only when he does, toward the novel's end. Third, the ostensibly “unmediated” access can allow a fuller exploration of changing and unusual minds because it puts the reader's focus on the consciousness and not the mediator (see PSYCHOLOGICAL).

Character Narrators and the First-Person Narrative Situation

A character narrator is defined by her participation in the narrative she actively mediates; or as Stanzel puts it in his description of the first-person narrative situation, the realms of existence of the character and the storyworld must be identical. However, the degree of a character narrator's participation can vary considerably, from narrators who mostly observe the primary action line like Conrad's Marlow in Lord Jim (1899—1900) to narrators who are the protagonists of the primary action line like Brontë's eponymous Jane Eyre. In addition, character narrators vary in terms of artistic control and self-consciousness (Sternberg, 2008) because authors motivate their narrators in different ways—i.e., they can choose from many possible explanations for why the narrator delivers the narration, or offer no explanation at all. So although all character narrators lead double lives as narrators and characters, as a narrating-I and an experiencing-I, authors may emphasize a character narrator's life as a character, her function qua narrator, or attend equally to both.

One challenge of character narration is that the author must communicate to the authorial audience through the character narrator's story to his narrative audience. From this perspective, character narrator is an “art of indirection” because the author must communicate indirectly through the limited perspective and realistic communicative frame of the character narrator's story to a dramatized or undramatized narratee that cannot be the actual reader (Phelan). Accordingly, rhetorical narratologist James Phelan has made an influential distinction between “disclosure functions” and “narrator functions.” A character narrator's disclosure functions involve the information of all kinds that the author wants to indirectly reveal to the actual reader. Narrator functions involve all the information that the narrator directly gives to her narratee. The value of this distinction is that it can explain why narrators sometimes offer their narratee information that he or she would presumably already know, what Phelan calls “redundant telling”: because the author needs to disclose that information to the authorial audience. Similarly, sometimes character narrators do not reveal their full relevant knowledge immediately, what Phelan calls “paradoxical paralepsis,” because the author needs to keep that information hidden, perhaps for plot tension. In short, the distinction between disclosure functions and telling functions helps readers consider the author's purposes for the character narrator's discourse.

As Wayne Booth first articulated, the personalization of character narrators especially evokes the question of (un)reliability: In what ways do the norms, values, and judgments of the narrator resemble or diverge from the implied author as recoverable from the narration? (Un)reliability remains a significant subject of study with respect to character narrators. Much debate focuses on the definition and utility of the implied author as a way of studying (un)reliability. For example, some critics define the implied author as a purely textual construct, others as a streamlined version of the real author. Still others argue that the implied author concept is not a useful way of understanding unreliable narration. In general, the implied author concept takes significant importance in the author-centric approaches that seek to know the author's intention (McCormick). This research has shown that character narrators can be (un)reliable with regard to their facts, interpretations, or judgments (Phelan, 50) and their reliability can change at different points in the narrative discourse. In contrast, reader-centric approaches emphasize that the hypothesis of unreliability is only one way that readers can account for anomalies in the text, especially if they don't center their reading on authorial intention (Yacobi). In general, unreliability studies intersect with many other questions, including the historical reception of texts, how readers make textual judgments, and author/reader relations, and so will likely continue to be a rich area of research in the future.

Just as (implied) authors have various relations with their character narrators, narrating-I's have various relations with those versions of themselves living in the primary-action level, the experiencing-I. Cohn (1978) makes a valuable distinction between “consonant” narrators, who identify with their experiencing-I, and “dissonant” narrators, who claim moral and intellectual distance from their former selves. For Cohn these categories can apply to external narration as well, but her terms are especially valuable to discuss character narration, and particularly when the experiencing-I is the protagonist of the primary-action level, as in J. D. Salinger's The Catcher in the Rye (1951) or Dickens's David Copperfield (1850).

Another important relationship involving character narrators is that between a narrator and the communities they may represent. Bakhtin discusses how the speech register of a particular character or narrator may represent an entire community of people who use the same kind of ideologically inflected language. Susan Lanser (1992) has shown that in novels like Sarah Orne Jewett's The Country of the Pointed Firs (1896), the narrator can situate herself inside the community she seeks to represent, and to some extent become an “I” that speaks for “we.” Such communal voices are interesting similarities and differences from those novels that actually use “we” to represent a particular community, like in Conrad's The Nigger of the Narcissus or the opening of Gustave Flaubert's Madame Bovary. (For more on “we” narration, see Richardson, 37—60; and Margolin, 591—618.)

Future Study

Recent studies of narrators have brought much needed attention to the use of simultaneous narration, camera-eye, and “unnatural” narrators of all types, including those in novels and short stories of dubious or limited narrativity. In addition, second-person narration is a particularly interesting case of “unnatural” narration because it does not fit cleanly into any of Stanzel's or Genette's categories (Fludernik 1996, 226; Richardson, 28). These are promising research topics for the future, as are studies of narrators in postcolonial novels and in different genres and media.

SEE ALSO: Editing, Formalism, Linguistics, Narrative Structure, Rhetoric and Figurative Language, Speech Act Theory, Story/Discourse.


1. Alber, J. (2009), “ Impossible storyworlds—and what to do with them,” Storyworlds 1: 79—96.

2. Bakhtin, M.M. (1981), Dialogic Imagination, ed. C. Emerson, and M. Holquist.

3. Banfield, A. (1982), Unspeakable Sentences.

4. Booth, W.C. (1983), Rhetoric of Fiction, 2nd ed.

5. Chatman, S. (1990), Coming to Terms.

6. Cohn, D. (1978), Transparent Minds.

7. Cohn, D. (1981), “ The Encirclement of Narrative,” Poetics Today 2: 157—82.

8. Culler, J. (2004), “ Omniscience,” Narrative 12(1): 22—34.

9. Fludernik, M. (1996), Towards a “Natural” Narratology.

10. Genette, G. (1980), Narrative Discourse.

11. Hamburger, K. (1973), Logic of Literature, trans. M.J. Rose.

12. Lanser, S.S. (1981), Narrative Act.

13. Lanser, S.S. (1992), Fictions of Authority.

14. Lubbock, P. (1921), Craft of Fiction.

15. Margolin, U. (2000), “ Telling in the Plural,” Poetics Today 21(3): 591—618.

16. McCormick, P. (2009), “Claims of Stable Identity and (Un)reliability in Dissonant Narration,” Poetics Today 30(2): 317—52.

17. McHale, B. (2008), “Ghosts and Monsters,” in Blackwell Companion to Narrative Theory, ed. J. Phelan and P. Rabinowitz.

18. Miller, D.A. (1981), Narrative and its Discontents.

19. Olson, B.K. (1997), Authorial Divinity in the Twentieth Century.

20. Phelan, J. (2005), Living to Tell About It.

21. Richardson, B. (2006), Unnatural Voices.

22. Stanzel, F.K. (1984), Theory of Narrative, trans. C. Goedsche.

23. Sternberg, M. (1978), Expositional Modes and Temporal Ordering in Fiction.

24. Sternberg, M. (1982), “ Proteus in Quotation-Land,” Poetics Today 3: 107—56.

25. Sternberg, M. (2008), “Self-Consciousness as a Narrative Feature,” in Blackwell Companion to Narrative Theory, ed. J. Phelan and P. Rabinowitz.

26. Walsh, R. (1997), “ Who is the Narrator?,” Poetics Today 18(4): 495—513.

27. Warhol, R. (1986), “ Toward a Theory of the Engaging Narrator,” Papers of the Modern Language Association 101(5): 811—18.

28. Warhol, R. (2008), “Neonarrative; or, How to Render the Unnarratable in Realist Fiction and Contemporary Film,” in Blackwell Companion to Narrative Theory, ed. J. Phelan, and P. Rabinowitz.

29. Yacobi, T. (2001), “ Package Deals in Fictional Narrative,” Narrative 9: 223—9.