Narrative Fiction before 1850: Instances of Refutation for Poetic Theories of Narration?^[1]

Sylvie Patron

University of Paris Diderot-Paris 7

1. In "Fiction, Pretense, and Narration", an article which has often been cited since its publication on the occasion of the French edition of Käte Hamburger's Die Logik der Dichtung (The Logic of Literature),^[2] Jean-Marie Schaeffer focuses on Hamburger's narrative theory, which he considers to be incompatible with narratology, or at least with certain aspects of narratology. In the last part of his article, he offers the following argument:

"[...] contrary to what Hamburger thinks, the ideal model that emerges from her analyses is not that of fictional narrative in the third person, but a much more specific model - that of heterodiegetic narrative with internal focalization, that is, with a narrator who recounts a story from which he is absent while adopting the perspective of his characters or character. Indeed, the definition of a fictional narrative as a Here and Now presentation of characters and their inner life only seems to be realizable via strict internal focalization, any element of external focalization resulting fatally in a split of the narrative universe (a split that Hamburger's theory does not recognize). In pure form, this model of narrative (which, as a global model, scarcely dates back to before the second half of the nineteenth century) seems quite distant in fact from serious narrative: the idea that description of the interior life of a third person is prohibited in serious narrative seems unassailable at first glance." (160)

I will single out one point only from this passage (notably leaving to one side the issue of whether the description of the inner life of a third person is prohibited in serious narrative or not).^[3] Hamburger put forward a theory which claimed to be a general theory of fictional narrative. According to Schaeffer, Hamburger's theory would not account for fictional narratives before 1850; in other words, the latter would refute it, or would at least refute some of its essential propositions. In this article I would like to examine the legitimacy of his assertion, in the case not only of Hamburger's theory, but also of theories by S.-Y. Kuroda and Ann Banfield, which I propose, following the latter two theorists, to call "poetic theories of narration".^[4]

As is clear from the passage quoted, like many other narratologists, Schaeffer is less interested in what Hamburger "thinks" or expresses explicitly, than in what "emerges" or is supposed to emerge from her analyses, which she neither thought nor expressed. However, it should be noted that not all narratologists have the same view of the model emerging from Hamburger's analyses. Dorrit Cohn writes, for example, that: "[...] starting out from textual observations, [Hamburger] demonstrates that certain language patterns are unique to fiction, and dependent on the presence of fictional minds within the text" (Transparent Minds 7) and also that:

"[...] where Hamburger's theory differs from all of these typologies [Norman Friedman's, Franz Stanzel's, Gérard Genette's, etc.] is that for her, psychic omniscience is not a narrative type or mode or device or technique, but the pivotal structural norm that rules the realm of third-person fiction and that is logically ruled out in all other discursive realms." (Distinction 25)

For Cohn, the model which emerges from Hamburger's analyses is not that of "heterodiegetic narrative with internal focalization", but that of the kind of narrative which is traditionally termed "omniscient", because it contains information unavailable to individuals with ordinary cognitive abilities (on the other hand, Cohn is right to say that, for Hamburger, the concept of psychic omniscience is not simply an issue of taxonomy).

It should also be noted that Schaeffer ("Fiction, Pretense" 162) is a relativist when it comes to Hamburger's theory ("this model of narrative [...] scarcely dates back to before the second half of the nineteenth century" and, later, "Once we look further than the historical field that the author of The Logic of Literature privileges [...] the phenomena of fictionalization are far from being so omnipresent as her theory leaves it understood to be").^[5] By contrast, he is hardly a relativist in his use of narratological categories: the opposition between "homodiegetic" and "heterodiegetic" narrative (163, 166), which supports the narratorial theory of narrative, in other words, the theory of a narrator in all narrative;^[6] "focalization" (167), defined as "the question of knowing from what viewpoint events are narrated" (despite the fact that narratology also use the term "focalization" to refer to "non-focalization" or the absence of point of view); the opposition between "internal focalization" and "external focalization". It is as though it were taken for granted that in any narrative with point of view, "a narrator [...] recounts a story from which he is absent while adopting the perspective of his characters or character" and that "any element of external focalization result[s] fatally in a split of the narrative universe". Yet Schaeffer may be considered, particularly with the benefit of hindsight, to be too ready to take for granted an issue which in fact begs further investigation.

The two points I have just cited make it possible to explain, in part, why Schaeffer pays only scant attention to the law which, according to Hamburger (83), governs third-person fictional narrative, to wit the disappearance of the author as the point of origin of the referential values and anchors of deixis ("I-Origo" in Hamburger's terminology), correlative with the possibility of representing "the I-originarity (or subjectivity) of a third-person qua third-person". Schaeffer ("Fiction, Pretense" 150) quotes Hamburger's famous statement regarding the use of verbs expressing internal processes in the third person ("She thus concludes that 'Epic fiction is the sole epistemological instance where the I-originarity (or subjectivity) of a third-person qua third person can be portrayed'"); he raises the issue on several occasions with reference to temporal and spatial deixis and to the epic preterit. However, he does not give the proposition any real theoretical status and, in particular, he does not discuss it in the final part of his article, which is devoted to a comparison between Hamburger's narrative theory and narratology. To my mind, by contrast, the proposition is at the heart of Hamburger's theory as well as Kuroda's and Banfield's. It is fundamental to the comparison which should be made between poetic theories of narration and narratology.

2. I will now briefly sketch a version of narratological theory in order to determine the extent to which Hamburger's, Kuroda's and Banfield's theories are compatible or incompatible with it.^[7] Narratology can be summarized by the following propositions:

(i) In any narrative, the story (the series of events told in the narrative) can be distinguished from the narrative itself (the spoken or written discourse which recounts the events). This distinction traditionally relies on the existence of discordances between the chronological order of the recounted events and the order of their disposition in the narrative.

(ii) The narrative is always recounted by somebody addressing somebody else (even in the case of written narrative: here, "recounted" signifies "produced in an oral or written verbal form"): this is narrating in Genette's terminology.

(iii) In the case of fictional narrative, the story and the narrating (and thus the narrator and the narratee) are fictional. More precisely, a fictional act of narrating duplicates the author's real act, which narratology does not analyse, although without it there would simply be no narrative.^[8] The fictional narrator recounts a series of events, with which he or she is familiar before recounting them, to the fictional narratee. It is the narrator who makes use of the categories of time (order, duration, frequency), mood and voice in Genettian narratology. It is the narrator who is responsible for the selection and presentation (sometimes termed "focalization", cf. Bal, Narratology  8 et passim) of narrative information in other narratologies.

(iv) Genette distinguishes between homodiegetic fictional narratives (whose narrator is present as a character in the story he or she recounts) and heterodiegetic fictional narratives (where the narrator is absent from the story recounted). However, this distinction is only a secondary distinction within the category of fictional narrative, defined as the discourse of a fictional narrator. Other narratologists make an equivalent distinction between fictional narratives narrated by an internal narrator or an external narrator (cf. Chatman 155, 158, 170 and Stanzel 4-5 et passim) and by an external narrator or a narrator-character (cf. Bal, Narratology 22 et passim, Chatman 234 and Stanzel 90).

(v) However, Genette makes no distinction between heterodiegetic fictional narratives in which the narrator refers to himself or herself using a first-person pronoun and those in which the narrator never refers to himself or herself and is not only absent as a character in the story recounted, but totally invisible in the narrative.^[9]

(vi) These propositions are not linked to any given narrative tradition. In any given tradition, individual author's work or any particular narrative, the importance given to particular elements of time, mood or voice may vary; by contrast, the story, the narrative and the narrating, the fictional nature of the story and its narrating in fictional narrative, as well as the distribution of fictional narratives into the categories of homodiegesis or heterodiegesis are all invariable.

(vii) The fact that the fictional narrator makes use of the categories of time, mood and voice, or the fact that the narrator is homodiegetic in certain fictional narratives and heterodiegetic in others, or again, that certain heterodiegetic narrators never refer to themselves, are not responsible for any special "semantic effects" which might differ from the contents and relations determined in (i) and (ii) and play a role in the interpretation of a particular fictional narrative. On this point, Genette's narratology differs from other narratologies which are more concerned with the issue of interpretation.^[10]

(viii) Narratology claims a similarity to John Searle's theory of fictional discourse (cf. Ryan "The Pragmatics of Personal" 518-19, Genette, Fiction 70-71, and Schaeffer, "Le récit fictif" 53, among others). It sees as equivalent Searle's proposition that "the author of a work of fiction pretends to perform a series of illocutionary acts of assertion" and the proposition which can be deduced from (iii), that "the fictional narrator performs illocutionary acts of assertion". In doing so, it extends the description reserved by Searle for the first-person fictional narrative, or a version of it, to all fictional narratives. I have tried elsewhere to show that this extension is abusive and indeed contradictory from Searle's perspective (cf. Patron, Le Narrateur 100-106, 123-25, 126-27, 130-33).

This presentation could give rise to long commentaries on the origins of such propositions, their formulation, which can vary from one author to another (certain propositions may even remain implicit) and on their retention by the school of narratology which is now called "post-classical".^[11] I will simply offer two comments.

The theory (i)-(viii) is a minimal, non-linguistic version of narratological theory. More content might be afforded to the theory through interpreting the propositions (ii), (iii), (iv) and (v) from a linguistic point of view. Insofar as it is a narratorial theory of narrative, narratology implies that any sentence of a narrative is considered as a message communicated by a narrator, whether real or fictional, to its receiver or narratee, whether real or fictional (the issue of knowing precisely what "fictional" means will be left to one side for the time being). For fictional narratives, one could speak in terms of double communication or communication on two levels, the first one real, the second fictional, by forming the hypothesis of a relation between the two levels similar to quoting direct speech without introduction or comment clause; however, narratology does not analyse the first level, so that the fictional level takes the place of the two levels by itself. The distinction between first-person fictional narratives (homodiegetic narratives, in Genette's terminology) and third-person fictional narratives (heterodiegetic narratives) has no pertinence within this communicational framework: any sentence in a fictional narrative is considered as a message which is communicated either by a character who refers to himself or herself as I, or by a speaker who can never refer to himself or herself, but is nevertheless still a first person who is not actual but can be actualized as such.

As I have just pointed out, the theory from (i) to (viii) contains a blind spot. Narratology affirms that, in fictional narrative, the narrator and the narratee are fictional, yet it has no concept of fictionality. It follows quite simply from its affirmations that it applies a double standard, one holding for the narrator of first-person fictional narratives, as well as the other characters, places, etc. in the fiction which are often termed "fictional entities" and are considered to have the same properties as entities in the real world, and another holding for the narrator of third-person fictional narratives (I am thinking in particular of those in which the narrator never refers to himself and is totally invisible and absent from his narrative). If the first narrator is undoubtedly fictional, the second might more legitimately be termed "postulated" (in the theoretical sense) or simply "theoretical".^[12]

3. I shall now turn to poetic theories of narration. It is much more difficult to present an acceptable version of these theories in a succinct manner than it is for narratology. It is also difficult to rectify the misreadings and misrepresentations to which they have given rise and potentially provide more content to elements which remain undetermined, without being able to discuss them at length (cf. Patron, Le Narrateur ch. 7-9). I shall limit myself to emphasizing three aspects: their calling into question of the idea of communication; their refutation of the narratorial hypothesis in certain precise cases and their reinterpretation of an old theory in new theories.

3.1. For the representatives of poetic theories of narration (precisely, Kuroda and Banfield), there is nothing obvious in the idea that the relation between author and reader in fictional narrative is one of communication in any essential or interesting sense of the term "communication". The adjectives "essential" and "interesting" are Noam Chomsky's (56, 57), quoted by Kuroda ("The Reformulated Theory" 69).^[13] In order to speak of communication in an interesting manner, it is necessary to be able to rely on a linguistic and, potentially, a pragmatic analysis of that which pertains to communication, as opposed to that which can be understood not to pertain to it. Neither is there anything obvious in the idea that there is a fictional communicational relationship in all fictional narratives which implies a fictional narrator and narratee (Kuroda and Banfield take the same position as Hamburger here).

Based on certain characteristics of the Japanese language (usage of the adjectival form of adjective/verb pair for expressing feeling, usage of the word zibun in certain structures), Kuroda's linguistic analyses establish:

(i) homogeneity between sentences in ordinary communication and sentences in first-person fictional narrative,^[14] as well as sentences in certain third-person fictional narratives which could be seen as narrated by a narrator who never refers to himself or herself (examples might be "neutral" or "behaviourist" narratives such as certain short stories by Hemingway);^[15]

(ii) heterogeneity, by contrast, between sentences in ordinary communication and certain sentences in certain third-person fictional narratives (third-person fictional narratives with the representation of the "point of view" of one or several characters: Joyce's Portrait of the Artist as a Young Man or Woolf's To the Lighthouse, for example).^[16]

Sentences which use the adjectival form of the adjective/verb pair for expressing feeling (kanasii instead of kanasigaru ["sad"]) or the word zibun ("self") for representing "point of view" or the "lived experience" (Erlebnis) of one or more characters in the third person cannot be described adequately within the framework of the communicational theory of linguistic performance which is based on the concepts of speaker and receiver. On the one hand, Kuroda states that:

"[t]he role of the writer of a non-first-person story is to assemble (in fact, create) such information and set it in order. This is in no way identifiable with the role of the "speaker" in the paradigm of linguistic performance [...]". ("Where Epistomology" 382-3)

On the other hand, he uses precise linguistic tests (insertion of yo at the end of the segment, rewriting of the sentence in the first person), to show that such sentences cannot be seen to be narrated by a fictional narrator adopting the viewpoint of one or more characters or taking on the task of presenting their Erlebnis.

Kuroda then refers to Benveniste on histoire and discours, Hamburger on fiktionales Erzählen and Banfield on free indirect discourse (represented speech and thought) to fuel his critique of communicational theories of narrative. He formulates Genette's revision of Benveniste's opposition between histoire and discours very clearly ("Reflections" 116, 125, 138).^[17] He also puts forward a new theory of linguistic performance to replace communicational theory and provide access to a unified conception of narrative and discourse in general. His theory is based on the identification and characterization of the objective function, as he puts it, of the sentence, as distinct from its communicative function. The distinction points to the fact that any sentence which has been created as a real entity in the world calls up a meaning in the mind of the person hearing or reading it - and does so whether the sentence is materialized in a situation of communication or not. Kuroda writes, for example, about the sentence containing zibun:

"Reading this sentence we obtain an image or knowledge of an event, but we do not assume the existence of any consciousness which has judged the occurrence of this event and communicated it to someone. Simply the sentence creates in us the image or knowledge of the event. This and this much is the function of the sentence vis-à-vis the reader." (134)

Based on the comparison between the effects produced by speech acts like orders or information (as distinct from assertions) on an outside listener and on the speaker's addressee in the strict sense, Kuroda's pragmatic analyses lead to the identification and characterization of three linguistic functions: the objective function (which is the most essential, in the sense that it is an essential component of the second and third functions); the objectifying function which consists, for a speaker, in objectifying the content of his mind (this is an essential component in the third function) and the communicative function which consists, for a speaker, in communicating the objectified content to the addressee (this it the least essential function in the sense that it presupposes the others but is not presupposed by them). A speech act is only an act of communication if the communicative function is activated. Kuroda indicates in passing that in Japanese, the communicative function can be recognized by specific formal traits (stylistic traits, terms and honorific constructions). He comes to the conclusion that the crucial characteristic of "the communicational theory of narration" (the quotation marks are in his text) is more closely related in the end to the objectifying function than the communicative one:

"The core of the theory is the assumed existence of a narrating agent. A narrative is assumed to be a product of the objectifying act of the narrator. Opinions concerning the role of the audience or the reader can vary." ("Some Thoughts" 10-11)

To this he opposes his own poetic theory of narration:

"The supposition of a narrator who objectifies the content of his consciousness in the form of a narrative is quite unnatural, especially in certain types of modern fiction where the inner experiences of multiple protagonists are simultaneously depicted. The function of language in such fiction cannot be accounted for in terms of the communicative or objectifying function. I have thus identified the function that makes such use of language possible as the objective function of language, the function such that an attentive consciousness cannot fail to respond to objectively materialized sentences and make out of them whatever sense can be constructed from them, without any necessity of recognizing or hypothesizing the existence behind those sentences of a consciousness objectifying what it perceives or judges." (11)

Banfield's linguistic analyses, to which I shall return in the next section, make use of Kuroda's and further them on a certain number of points: bringing the axiomatic nature of communication in the communicational theory of narration into question; affirming the necessity of defining communication in rigorous linguistic terms; defining linguistic subjectivity (this definition is based on a comparison between direct discourse, indirect discourse and free indirect discourse, or represented speech and thought, which enables the formal marks of the expression of subjectivity and those of communication to be isolated); introducing, alongside the idea of the speaker, the new idea of the subject of consciousness or SELF to designate the origin of deixis and subjectivity; characterization of sentences in third-person free indirect discourse as sentences which are not governed by the framework structured by the relation of communication between an I and a you.

Reflection on the role of writing can also be found in Banfield's work, in the sense of written linguistic realization (as distinct from the transcription of oral language) in the production and reception of fictional narratives. In Banfield's theory, writing is the extra-linguistic factor which enables certain potentialities of language to be actualized in performance.

3.2. This section is devoted to the refutation of the narratorial hypothesis in several specific cases. Banfield offers both the most powerful counter-examples and the most refined theory enabling explanation both of counter-examples to the initial hypothesis and of the facts it explained. I shall simply recall the following points:

(i) the principles she formulated for free indirect discourse (represented speech and thought): the principle of the uniqueness of the subject of consciousness, according to which "[f]or every node E, there is at most one referent, called the 'subject of consciousness' or SELF, to whom all expressive elements are attributed"^[18] (this principle is a reformulation of the "one expression/one speaker" principle for direct and indirect discourse); the principle of the priority of the speaker, according to which "[i]f there is an I, I is coreferential with the SELF" (Unspeakable 93).

(ii) the test enabling these principles to be justified: if an I is added to a sentence of third-person free indirect discourse, it becomes clear that it is no longer possible to attribute expressive elements and constructions to a subject other than the referent of the I (in other words, it is no longer possible to consider them as sentences of third-person free indirect discourse):

"Where were her paints, she wondered? Her paints, yes."  (Woolf, To the Lighthouse 168).

Where were my paints, *she wondered? My paints, yes.^[19]

As Banfield puts it:

"Since no first person may appear in represented speech and thought except one interpretable as the E's SELF and since that first person must also appear in any parenthetical attached to the represented E, this means that represented Es cannot be simultaneously attributed to a covert or 'effaced' narrator." (Unspeakable 97)^[20]

Note that in Banfield's view, if sentences of free indirect discourse do not necessarily have a narrator, they do have an author who is responsible for their existence as real entities in the world as well as their disposition and role within the text.

The same test could be used and the same reasoning put forward in relation to sentences which, according to Banfield, represent subjectless subjectivity:

"The sun had now sunk lower in the sky." (Woolf, The Waves 129)^[21]

It is clear that these sentences cannot be attributed to a covert or effaced narrator (or one revealed indirectly by the use of the adverb "now") without losing their essential characteristics: the combined occurrence of "now" and a past tense; the absence of any subject of consciousness to whom the "now" might be attributed.

3.3. This section concerns the question of the reinterpretation of an old theory within new theories. I shall only deal with Hamburger's theory, but my comments can be extrapolated to Kuroda's and Banfield's.

The conceptualization of the difference between third-person fictional narrative, in which the author himself recounts the story and first-person fictional narrative (in the form of an embedded narrative^[22]), in which a character fictionally recounts the story, dates back to Aristotle. The concept of a narrator distinct from the author made its appearance much later, in the early 19th century, in the first descriptions of first-person fictional narrative without embedding (fictional memoirs or first-person novels in the proper sense).^[23] The issues it encapsulates are the following: an I which does not belong to the author but to a fictional character; a pact of truthfulness within the fiction (the narrator speaks of the past in a truthful manner, which does not mean that he cannot lie or deform the truth, but simply that he does not relate a fictional narrative); restriction of narrative information to knowledge open to the I and to his memory; potentially a restriction of thought and style proper to the ability of the I; some degree of contrast in terms of their age, social status, general knowledge and rational ability between the narrated I in the story and the I narrating the story.

In Hamburger's theory, the following reappear:

- the difference between third-person fictional narrative, considered as a general or standard case and first-person fictional narrative, considered as a particular case;^[24]

- the traditional view of the narrator reserved for first-person fictional narrative: Hamburger claims that "only in cases where the narrative poet actually does 'create' a narrator, namely the first-person narrator of the first-person narrative, can one speak of the latter as a (fictive) narrator" (140);^[25] she thus rejects the Stanzelian view of the (fictitious or fictional) narrator, extending to third-person fictional narratives with intrusions such as "I", "we", "our hero" and, from there, to all third-person fictional narratives;^[26]

- the issues implied by the traditional concept of the narrator, in particular the issue of the narrator belonging to the fictional world and the restriction of information to knowledge (of the inner life of the other characters) and memory available to the narrator.

By contrast, Hamburger is the first theoretician to conceptualize:

- the fact that the author of a third-person fictional narrative does not narrate in the same way as the author of a factual narrative or the narrator of a first-person fictional narrative within the fiction: he or she appears to narrate events which exist prior to the act of narration (which Hamburger expresses by provisionally retaining the term of "narrator") (cf. 62 et sq.),^[27] but in reality, he or she "creates" such events, or makes them exist fictionally (which Hamburger conceptualizes in terms of "functional correspondence" between the narration and that which is narrated (cf. 136), while the narrator of a first-person fictional narrative "does not 'engender' that which he narrates, but narrates about it in the same manner as in every reality statement [...]" (317) (which Hamburger conceptualizes in terms of "feigned reality statements") (cf. 311, 313 et sq.);

- the fact that the author of a third-person fictional narrative disappears as the point of origin for referential values and deictic markers (real "I-Origo"), which allows the apparition of "fictive I-Origines" whose "I-originarity (or subjectivity)" may be developed to a greater or lesser degree (cf. 73 and 83),^[28] while the narrator of a first-person fictional narrative "is always present" as I-Origo of the narrative and "never disappears, which [...] would result in the emergence of fictive I-origines in its place" (315);^[29]

- the fact that the third-person fictional narrative bears linguistic and textual signs of this appearance/disappearance: verbs describing inner processes in the third person, monologue and free indirect discourse, past tenses combined with present and future deictic adverbs, situational verbs combined with indications of date and long passages of dialogue (the list should not be considered exhaustive); in principle, such signs should not be used in first-person fictional narratives; in practice, they are often found combined with signs of first-person narration (few authors renounce long passages of dialogue, for example) (cf. 332, 337);

- the fact that third-person fictional narrative is a narrative with no narrator, since the hypothesis of an omniscient narrator as a corollary to that of a fictional narrator in third-person fictional narratives is so vague and devoid of independent justification that it can only be viewed as a pseudo-hypothesis (cf. 141).^[30]

I have two final comments to make.

Hamburger's theory does not deny those aspects of fictional narrative which were examined by narratology under the category of time (order, duration, frequency).^[31] However, her theory, like those of Kuroda and Banfield, attributes the use of such categories, or techniques of composition, not to the narrator, but to the author - as was current before Genette and the advent of narratology. Their theories distinguish elements pertaining to fictional content (characters, places, etc. and the narrator if there is one) from the means put to work in constructing the fiction (language, style, techniques of composition).^[32]

Searle's and Hamburger's theories are not fundamentally incompatible.^[33] Notably, they share elements stemming from the traditional view: the difference between third-person fictional narrative and first-person fictional narrative, the view of the narrator, etc. They also share a study of how reference works in fictional narrative (in Speech Acts, even evokes sentences which cannot be written or said: "Sherlock Holmes is coming to my house for dinner tonight" (78), for the reference to "my house" puts us back in real world talk). It is true that Searle claims categorically that "[t]here is no textual property, syntactical or semantic, that will identify a text as a work of fiction" ("The Logical Status" 65). But the necessary and sufficient properties for identifying a text as a work of fiction (Searle) should be distinguished from signs pointing to the fictional character of the existence of that which is narrated (Hamburger).^[34] Besides, if Searle had paid closer attention and taken his reading further, he would have found numerous signs of fictionality in Hamburger's sense in his chosen example of a fictional work (Iris Murdoch's The Red and the Green), some of which could have been considered sufficient to identify the text as a work of fiction (cf. Wildekamp et al., notably 559-561, 565). As for Hamburger's theory, she does not deny that many sentences or passages in fictional narratives are formally identical to those found in either oral or written factual narratives.^[35]

4. I mentioned at the outset that Hamburger put forward a theory which claimed to be a general theory of fictional narrative. However, her generalizing aim is not achieved in the same way in all the domains she examines. In the case of signs of fictionality, for example, globally speaking the theory attains a certain degree of generality - Hamburger does not claim that all the signs discovered and explained are found in all third-person fictional narratives, whatever their genre, period or the author under consideration.^[36] By contrast, Hamburger does put forward a theory of the history of third-person fictional narratives (essentially incarnated in the novel as a genre), which is that of the ever greater autonomy in the presentation of characters in relation to that which a real subject could say or write of people belonging to a shared reality (this is what she understands by forming characters "as fictive persons or subjects") (59). She writes, for example:

"We already have pointed out on several occasions that in the course of the nineteenth century the techniques of fictionalization had become more and more refined, the presentation of the psychic life came increasingly more to employ the devices of directly subjectifying the characters. That is the fictive I-originarity of the figures became more and more explicitly developed, culminating in the bold methods of Joyce." (168)

In the same way, Kuroda's and Banfield's theories aim to be theories of fictional narrative and not of a given corpus of fictional narratives, even if this may appear less obvious than in Hamburger's case. Remember the sentence from Kuroda ("Some Thoughts" 11) quoted above:

"The supposition of a narrator who objectifies the content of his consciousness in the form of a narrative is quite unnatural, especially in certain types of modern fiction where the inner experiences of multiple protagonists are simultaneously depicted." (my italics)

I do not have time to present the "deictic shift theory" developed by a group of researchers in cognitive science at the State University of New York at Buffalo, which forms an attempt at generalization, as well as an empirical validation, of poetic theories of narration (cf. Duchan et al., notably ch. 1, 2 and 6).

Is it the case that fictional narratives prior to 1850 are not accounted for by poetic theories of narration - in other words, that they refute them, or at least refute some of their essential propositions? For this to be demonstrated, it would first need to be shown that a particular fictional narrative prior to 1850, chosen for a particular emblematic reason, formed an authentic counter-example to the theories or to some of their essential propositions. Next, multiple analyses of as wide and varied a range as possible of pre-1850 fictional narratives would need to be carried out. Finally, as part of the comparison between narratology and poetic theories of narration, it should be ascertained that the theory which esteems the counter-examples to be recognized as such is indeed narratology (one of the essential propositions of which, to wit "fictional narrative is the discourse of a fictional narrator", would need to be considered unrefuted). To the extent that this programme of work has not yet been carried out, I shall conclude by saying that if poetic theories of narration may perhaps be refutable by fictional narratives prior to the latter half of the 19^th century, we cannot yet consider them refuted.

Translated by Susan Nicholls

---

---

References