This article, translated here for the first time, develops some of the central arguments of Sylvie Patron’s books, The Narrator: A Problem in Narrative Theory (2023 [2009, 2016]) and The Death of the Narrator and Other Essays (2019 [2015]), and the collection Optional-Narrator Theory: Principles, Perspectives, Proposals (2021).^[1] 

Narratology (literally: the science of narrative) is usually regarded as one of the most developed sectors of literary theory. It is not uncommon for scholars to invoke its advanced methodology, its precise categories of analysis, the discriminating power of its terminology—even when it comes to theorizing or analyzing literary genres that do not a priori fall within its field of application.^[2]

But there is another discourse on narratology, less widespread, according to which, in this domain of theorizing, “there is now controversy on almost everything”.^[3] I would be inclined to adopt this second discourse instead and I myself contributed to the controversy in my work on the concept of the narrator.^[4]

In this article, I would like to return to the close association, which goes as far as identification, between narratology and the theory of the existence of a narrator in all narratives. More specifically, this theory postulates the existence of a narrator who merges with the author in all non-fictional narratives (historical narratives, biographies, autobiographies, etc.) and the existence of a fictional narrator different from the author in all fictional narratives (novels, short stories, tales, etc.). I will refer to this theory as pan-narrator theory and I will focus exclusively on the version of the theory that concerns fictional narratives.

In the first part, I will present the essential aspects of the association between narratology and the pan-narrator theory of fictional narratives. In the second part, I will make a few remarks on the context of the emergence, then of the gradual establishment, of the theory, which may, it seems to me, have a more general relevance for the various sectors of literary theory.

1. Narratology and the Pan-Narrator Theory of Fictional Narratives

Roland Barthes had already suggested, in some programmatic remarks, that the scientific study of narrative, according to the structuralist paradigm, could and should only be concerned with the narrator, explicitly differentiated from the author.^[5] But it was Gérard Genette, a disciple of Barthes, who laid the foundations of the association tending to become ever closer between narratology and the pan-narratory theory of fictional narratives.

In “Discours du récit: Essai de méthode” (Figures III, 1972 [English: Narrative Discourse: an Essay in Method, trans.  Jane E. Lewin, 1980]), Genette begins by technically redefining the terms récit (narrative), histoire (story, the eventful content of the narrative) and narration (narrating or narration, the act that produces the narrative, anchored in a narrative situation). He states that “Analysis of narrative discourse will thus be […], essentially, a study of the relationships between narrative and story, between narrative and narrating, and (to the extent that they are inscribed in the narrative discourse) between story and narrating.”^[6] Of the difference between non-fictional and fictional narratives, he first accounts in a rather expeditious manner, saying that in one case story and narrating are real and can be attested by documents external to the narrative, while, in the other, they are fictional (Genette’s word is fictifs, “fictive,” used in the sense of “fictional”) and without external attestation: “It is thus the narrative, and that alone, that informs us here, both of the events that it recounts and of the activity that supposedly gave birth to it.”^[7] This difference is occasionally evoked again in a passage of the chapter entitled Voix (“Voice”), which specifically distinguishes the narrator of fictional narratives as being himself a fictive role, even if assumed directly by the author (in the English translation, with a slight difference: “where the role of the narrator is itself fictive, even if assumed directly by the author”^[8])—which falls under another kind of explanation of fiction.

In the chapter “Voice,” Genette posits that all narratives can be divided into two categories that are mutually and necessarily opposed: heterodiegetic narratives, whose narrator is absent from the story he tells, and homodiegetic narratives, whose narrator is present as a character in the story he tells.^[9] All the examples given refer to fictional narratives: “Homer in the Iliad, or Flaubert in L'Éducation sentimentale” and “Gil Blas, or Wuthering Heights”^[10]—with an ambiguity between, on the one hand, the names of heterodiegetic narrators, derived from the names of the authors, and on the other hand, the titles of homodiegetic narratives. The fact that the names of the heterodiegetic narrators are derived from the names of the authors of the narratives in question contradicts an earlier passage in the chapter “Voice,” in which Genette attributes to the narrator of third-person fictional narratives the property of being fictional, in the same way and ultimately for the same reason as the narrator of first-person fictional narratives: “[…] the references in Tristram Shandy to the situation of writing speak to the (fictive) act of Tristram and not the (real) one of Sterne,” but, Genette adds, “in a more subtle and also more radical way, the narrator of Père Goriot ‘is’ not Balzac, even if here and there he expresses Balzac’s opinions, for this narrator-author is someone who ‘knows’ the Vauquer boardinghouse, its landlady and its lodgers, whereas all Balzac himself does is imagine them…”^[11] A few lines later, Genette puts the name of “Homer” in quotation marks (“most of the Odyssey is told by ‘Homer’”), meaning: the fictional version of Homer that tells the Odyssey from within the fictional world.

The narrator goes hand in hand with the narratee, at the other pole of the diagram of the narrative communication: “Like the narrator, the narratee is one of the elements of the narrating situation, and he is necessarily located at the same diegetic level; that is, he does not merge a priori with the reader (even an implied reader) any more that the narrator necessarily merges with the author”.^[12]

Following Genette, all works claiming to belong to narratology adopted the main articles of faith of the pan-narrator theory of fictional narratives, of which they contributed to fix the formulas, still partially floating in Genette (“fictive,” “fictional,” “Homer” with or without quotation marks, etc.). The following quotation, taken from Israeli narratologist Shlomith Rimmon-Kenan, can be considered the clearest and most explicit formulation of the pan-narrator theory of fictional narratives:

Narration can be considered as both real and fictional. In the empirical world, the author is the agent responsible for the production of the narrative and for its communication. The empirical process of communication, however, is less relevant to the poetics of narrative fiction than its counterpart within the text. Within the text, communication involves a fictional narrator transmitting a narrative to a fictional narrated.^[13]

Genette himself considerably strengthened the association between narratology and the pan-narrator theory of fictional narratives, to the point of affirming their inseparable fusion, in his later work.

2. Some Remarks on the Context of the Emergence and Establishment of the Pan-Narrator Theory of Fictional Narratives

It may have been noted, in the brief presentation above, the scarcity of the attempts to justify, to demonstrate, the propositions put forward. These are falsely obvious, which can give the impression that they are natural. However, they are not. There is nothing natural in using the term “narrator” both for the character, that is to say the fictional person, man or woman, who writes or orally narrates his or her story, possibly to one or more other characters, in first-person fictional narratives (this is the traditional conception of the narrator, linked to the first theorizations of the memoir-novel or first-person novel in the traditional sense of the term), and for the “role” or the “instance” that, on the one hand, is supposed to narrate all narratives and, on the other hand, is supposed to be fictional in all fictional narratives. In the latter case, the term “narrator” refers to a theoretical entity, its referent is an abstractum, a construction, while the narrator of the traditional conception is an immediate empirical object: “For a long time, I used to go to bed early” (Proust, Swann’s way), “My father, Blas of Santillane, after having borne arms for a long time in the Spanish service, retired to his native place” (Lesage, Gil Blas), “My true name is so well known in the records or registers at Newgate, and in the Old Bailey, and there are some things of such consequence still depending there, relating to my particular conduct, that it is not to be expected I should set my name or the account of my family to this work [...]” (Defoe, Moll Flanders). (It is true that the authors can make the male or female narrator less immediate and defer the revelation of his or her identity, or even his or her existence in the story, but this possibility does not in any way detract from the reality of the difference between the narrator of the traditional conception and the narrator of pan-narrator theory.) It can also be said that the narrator of the traditional conception is a creation of the author, while the supposedly fictional narrator in all fictional narratives is a creation of a theory.^[14]

In the following sections, I will consider successively the lack of historical background of narratologists, the problematic and paradoxical nature of their relationship to linguistics, finally the various reformulations, corrections, adjustments made to the theory to allow it to integrate new theoretical proposals concerning fiction, presented in translation in the French context—as so many conditions participating in the specificity of the context of emergence and establishment of the pan-narrator theory of fictional narratives.

2.1. This section concerns the lack of historical background of narratologists or, to put it in more technical and more precise terms, their limited “horizon of retrospection,” centered on a synchrony barely extended to some antecedent references.^[15]

Genette gives no bibliographic reference for his use of the term “narrator.” In the introduction to Narrative Discourse he quotes Tzvetan Todorov, taking up the term as if it were obvious in this use (the discussion is not about the narrator, but about the name and extension of the categories of analysis): 

My starting point will be the division put forth in 1966 by Tzvetan Todorov. This division classed the problems of narrative in three categories: that of tense […]; that of aspect, "or the way in which the story is perceived by the narrator"; that of mood, that is, the "type of discourse used by the narrator."^[16] 

In the chapter “Voice,” the proposition “the narrator of fictional narratives is a fictional role” sounds like the echo of a passage from an article by the German theorist Wolfgang Kayser.^[17] The argument that the narrator of Father Goriot is someone who “knows” the Vauquer boardinghouse, etc., can also be considered as coming from Kayser’s article.^[18] Genette explicitly refers to Barthes for his use of the term “narratee,” adding a more unexpected reference to A. J. Greimas (concerning the couple that serves as a model for the narrator-narratee couple, namely the addressor-addressee).^[19]

Clarification on the emergence and application of the term “narrator” is therefore lacking, and no quotations or cautions against retrospective illusion are included in Genette’s Essay in Method.

Genette also makes historical mistakes, such as when he says at the opening of the “Person” section, 

Readers may have noticed that until now we have used the terms ‘first-person—or third-person—narrative’ only when paired with quotation marks of protest. Indeed, these common locutions seem to me inadequate, in that they stress variation in the element of the narrative situation that is in fact invariant— to wit, the presence (explicit or implicit) of the ‘person’ of the narrator. This presence is invariant because the narrator can be in his narrative (like every subject of an enunciating in his enunciated statement) only in the ‘first person’ […].^[20]

This paragraph prepares and justifies in advance the new terminological proposal that introduces the terms “homodiegetic” and “heterodiegetic.” The problem is that in the traditional definition of the terms “first-person” and “third-person narrative” (exactly “first-person” and “third-person novel”), the thematized element is not the narrator, but the hero. A first-person narrative (novel) is a narrative in which the hero is the narrator. A third-person narrative (novel) is a narrative in which the hero is a third person whose adventures we are told by the author.^[21] Contrary to what Genette writes, there is therefore nothing inadequate in the terms “first-person” and “third-person.”

Genette makes only a discrete and ambiguous allusion, at the beginning of the chapter “Voice,” in the section entitled “Time of the Narrating,” to the Logik der Dichtung of Käte Hamburger (1957, 2nd ed. 1968 [English: The Logic of Literature, trans. Marilyn J. Rose, 1993]), which is however an essential milestone in the history of the concept of the narrator.^[22] Hamburger is not mentioned in the section “The Narrating Instance,” nor in the section “Person,” where it should have had all its place, in a position of an opponent or challenger. To put it succinctly, Hamburger’s theory challenges the postulate of the fictionality of the narrator and the narrating situation in all fictional narratives, since it is a condition that it specifically reserves for first-person fictional narratives. It is equally incompatible with the assumptions of the distribution of fictional narratives into homo- and heterodiegetic.

The very limited nature of Genette’s horizon of retrospection, and this is also true of the narratologists who will come after him and who will refer essentially to him, is obvious in the following statement: 

Without a narrating act, therefore, there is no statement, and sometimes even no narrative content. So it is surprising that until now the theory of narrative has been so little concerned with the problems of narrative enunciating, concentrating almost all its attention on the statement and its content, as though it were completely secondary, for example, that the adventures of Ulysses should be recounted sometimes by Homer and sometimes by Ulysses himself.^[23]

The “theory of narrative” that Genette evokes here coincides strictly with the structural analysis of narrative, which is hardly more than eight years old at the time (going back to the first works of Claude Bremond)—as if no attention had been paid to the problems of narrative enunciating between Plato (quoted in the next sentence) and Genette, or as if the manifestations of this attention were unworthy of the qualification of theory. In both cases, it would be possible to show that these are not serious assumptions about the history of the discipline.

2.2. The problematic and paradoxical relationship of narratologists to linguistics is revealed in the fact that they insistently claim to the linguistic model, in its structural version inherited from Saussure (via a number of mediations), but it is to apply it to the nonlinguistic aspects of narratives. On the other hand, they ignore or exclude, more or less consciously, more or less explicitly, the proper linguistic approaches to their object of study.

According to Jean-Louis Chiss, “[t]he structuralist scientificity, originating from structural linguistics and the reactivation by Barthes of the Saussurian project of a general semiology, invested the French literary studies on the basis of conceptual transfers operating with analogy, homology, metaphorization.”^[24] Chiss gives the example of the term “language” in “the language of the narrative” (French, la langue du récit) in Barthes, that of the homological relationship between sentence and narrative in the same Barthes, that of the inflation of the term “grammar” in structural enterprises—but one could also invoke the term “discourse” in Genette’s “narrative discourse” (which in particular forecloses the question of the difference between orality and writing), the homological relationship between verb and narrative in Todorov and Genette, the genettian narratology as an analysis of tense, mood and voice in narrative (with an ambiguity in the use of the word “voice,” which, in tense, mood and voice, refers to the distinction between active and passive voice, but which is used by Genette as a synonym of “enunciating”). Genette himself speaks for the choice of these terms of “a kind of linguistic metaphor that should certainly not be taken too literally”^[25] and adds, following the introduction of the term “voice”: “once more, these terms are merely borrowed, and I make no pretense of basing them on rigorous homologies.”^[26]

On the other hand, the more or less conscious exclusion of language and linguistic approach can be illustrated by the following examples:

—  In the chapter “Mood,” Genette introduces a distinction between “Who speaks?” and “Who sees?,” between “person” (re-semantized in terms of the narrator’s relationship to the story he tells) and “focalization.” This distinction ignores the fact that a number of linguistic marks of focalization, in the case of “internal focalization” (what is traditionally called “point of view”) are largely the same as those of “voice”—apart from the first-person personal pronoun, “I,” which has a particular functioning. This is the whole subject of Hamburger’s reflection on what can be called the “character-related deixis,” a reflection that will be taken up and developed, in a work of greater linguistic technicality, by Ann Banfield. (As an example, the sentence "Tomorrow was Christmas," in which the deictic adverb "tomorrow" is attributed not to a "I" but to a character referred to in the third person, is quoted by Hamburger as a characteristic sentence of third-person fictional narratives.)

—  In the chapter “Voice,” Genette replaces the linguistic distinction, or at least recognizable and comprehensible by linguistics, between first- and third-person narratives, by a non-linguistic distinction, based on the narrator’s relationship to the story he tells (homo- or heterodiegetic). For Genette, the real question is not whether or not there are first-person marks, “[t]he real question is whether or not the narrator can use the first person to designate one of his characters.”^[27] Without saying that the question formulated by Genette is uninteresting (even if it should be reformulated: the real question is whether or not the author can use the first person to create a character who so fictionally designates himself or herself), the foreclosure of the distinction between narratives with and without first-person linguistic marks is indicative of a very distanced attitude towards linguistics.

—  In the same chapter, Genette formulates pseudo-evidences concerning “the presence (explicit or implicit) of the ‘person’ of the narrator […] [who] can be in his narrative (like every subject of an enunciating in his enunciated statement) only in the ‘first person’” or the fact that “the narrator can at any instant intervene as such in the narrative,” and that “every narrati[ve] is, by definition, to all intents and purposes presented in the first person.”^[28]

The history of the relationship between narratology and linguistics also includes an even less glorious moment, which is the one where Genette, in Nouveau discours du récit (English: Narrative Discourse Revisited) eleven years after “Discours du récit” (Narrative Discourse), endeavors to consider Ann Banfield’s theory from the perspective of what he calls “voice” or “person.” False claims (“Banfield’s starting point is the sound [if not original] observation that certain characteristic forms of the written narrative, like the aorist [the French passé simple] and free indirect speech, are by and large unknown in spoken language”); misinterpretations of certain terms (“From that exclusion in fact she derives an impossibility in principle: such sentences she alleges to be radically ‘unspeakable’”); generalizations and amalgams (“That sliding is characteristic of generative grammar, always quick to declare ‘unacceptable’ whatever has not yet been accepted”); word play (“So no one is speaking in them, and that is why your daughter is mute”^[29]), etc.: everything in these pages belongs to the pamphlet genre, aimed at diminishing and ridiculing the opponent, not to the argued scientific debate.

The works of Banfield but also those of S.-Y. Kuroda^[30] provide some tests to eliminate the hypothesis of the presence of an implicit narrator (without observable linguistic marks) in certain sentences of fictional narrative, in Japanese (Kuroda) or in English and French (Banfield). These sentences can generally be referred to as “sentences of free indirect style” (to use Charles Bally’s expression, even if it has taken on a broader meaning, with very varied definitions and fields of application) or “of free indirect style in the third person and the past tense.” A perhaps more operative description here would be “sentences that represent third-person subjectivity.” (We can quote again “Tomorrow was Christmas” or this sentence from Virginia Woolf’s Mrs. Dalloway: “He was in love! Not with her! A younger woman, of course!”, where “she” refers to Clarissa Dalloway). The proposed linguistic tests show that no personal pronoun or any other first-person linguistic mark, assumed implicit, can be made explicit without changing the meaning of the sentence. (In "He was in love! Not with her! Of a younger woman, I know!", the exclamations can no longer be attributed to the referent of "she," but necessarily to that of the first-person pronoun, "I", which functions as a kind of attractor for linguistic marks of subjectivity.)

In the case of sentences that represent third-person subjectivity, Genette’s proposals regarding “the presence (explicit or implicit) of the ‘person’ of the narrator [...] in the first person” or the fact that “[i]nsofar as the narrator can at any instant intervene as such in the narrative, every narrati[ve] is, by definition, […] in the first person,” these propositions must be considered refuted.

2.3. I will go more quickly through the various reformulations, corrections, adjustments made to the theory to allow it to integrate new theoretical proposals, presented in translation in the French context, then with a delay. These propositions fill a gap, which corresponds to something under-theorized, in narratology, concerning the difference between the act of the author who writes a non-fictional narrative and that of the author (who may be the same person as the first) who writes a fictional narrative.

In the preface to the French translation of Käte Hamburger (Logique des genres littéraires, 1986), Genette reaffirms the pan-narrator theory of fictional narratives to the point of absurdity: “[...] the work of fictional narratology, always more or less focused on the comparison of discourse and story, assumes (by virtue of a provisional methodological decision) that the nonserious pretense of fiction – to tell a story that has actually happened – is taken seriously”; 

[...] one cannot simultaneously study fictional narrative as narrative and as fiction, they remain disconnected: the ‘as narrative’ of narratology implies by definition that one pretends to accept (the fiction of) the existence, prio to the telling, of a story to be told […].^[31] 

Absurdity, because one can actually study fictional narrative both as narrative and as fiction—it seems to me that this is precisely what Hamburger does; but also because narratology here seems doomed to fictional illusion (and this despite the distancing marks: “by virtue of a provisional methodological decision”). The study of the fictional narrative “as fiction,” to take up Genette’s opposition, seems to be a more promising option for literary theory

In Fiction et diction (1991 [English: Fiction and Diction, trans. Catherine Porter, 1993]), which is composed of articles previously published in journals, Genette juxtaposes two contradictory interpretations of John R. Searle’s theory of fictional discourse. The first in the book (which is the second chronologically) seems to acknowledge the fact that Searle’s theory reserves a specific treatment to the first-person fictional narrative. 

[I]n the type of narrative called ‘personal,’ or ‘first-person’ (or, in more narratological terms, narrative with a homodiegetic narrator), the enunciator of the narrative, herself a character of the story (this is the only relevant sense of the expression ‘in the first person’), is herself fictional, and therefore the speech acts she performs as narrator are as fictionally serious as those of the other characters in her narrative […]^[32]

—“The only task that remains, then, is to describe the pragmatic status of impersonal or third-person narrative, […] heterodiegetic.”^[33]What interests me here, independently of the reaffirmation of the pan-narrator theory through the use of the terms “homo-” and “heterodiegetic,” is the fact that Genette does not extend to third-person fictional narratives the description that Searle specifically reserves for first-person fictional narratives. In the second interpretation of Searle’s theory (the first chronologically), Genette states that narratology and Searle’s theory are simply notational variants: 

to say, as Searle does, that the author (for example, Balzac) does not answer seriously for the assertions of his narrative (for example, the existence of Eugène Rastignac), or to say that we have to relate them to some implicit function or agency distinct from the author (the narrator of Le Père Goriot), is to say the same thing in two different ways; the choice between them is made on the basis of the principle of economy alone, according to the needs of the moment.^[34]

Genette extends here to third-person fictional narratives the description reserved by Searle for first-person fictional narratives, which is abusive and indeed contradictory from Searle’s perspective. 

I take from two Swedish theorists, Greger Andersson and Tommy Sandberg, a last remark that seems to me to have its place in this set. Andersson and Sandberg observe that narratologists present relevant and often excellent analyses and interpretations of many fictional narratives, which can only be explained on the assumption that they do not apply their theory rigorously. (This theory should indeed lead to strange interpretations, even aberrations, such as that which consists in trying to answer the question: How can the narrator of third-person fictional narratives “know” in the sense of Genette, or authoritatively narrate, the most intimate thoughts and feelings of the characters of his or her, or its, narrative?) Andersson and Sandberg rely on examples borrowed from great representatives of contemporary narratology, such as Liesbeth Korthals Altes and James Phelan. What they call “similarity theory,” abbreviation for “similarity theory between non-fictional and fictional narratives,” is closely related to what I call pan-narrator theory of fictional narratives.^[35] Their suggestion that it would be preferable to revise the theory commonly adopted by narratologists so that the theoretical statements are congruent with analytical and interpretative practices also applies to the pan-narrator theory of fictional narrative.

Conclusion

This article has revealed the fragile, problematic compactness of certain concepts and categories of narratology: narrator, voice (more accessorily mood or focalization), fictionality, homo- and heterodiegetic. The question now is whether it is possible to disassociate narratology and the pan-narrator theory of fictional narratives – in other words, whether narratology is capable of renouncing an outdated and even refuted theory.

The second part also implicitly contains a reflection on the conditions of scientificity of narratology. Among them, a serious relationship to the history of concepts and more generally to the history of the discipline or field of research. Narratology does not escape the condition formulated by Sylvain Auroux for the humanities, namely that in these disciplines, “historical inquiry must compensate for an often impossible experimentation.”^[36] Another condition: the need to ensure that what is called linguistics (or any other discipline) is recognizable by linguists (or any other representatives of the discipline in question), which seems a basic condition for any interdisciplinary dialogue.