The Logic of Literature, 2d rev. ed - Review

Kate Hamburger. Revised, 1968. Trans. Marilyn J. Rose. Bloomington: Indiana University Press, 1973; rpt. 1993. $35.00 cloth; $14.95 paper.

1

It is never too late to do something good: take for example the availability, in French translation, nearly thirty years after its first publication (1957), of one of the great works of contemporary poetics. In choosing to translate the original title, Die Logik der Dichtung, as Logique des genres litteraires, the author of this excellent translation has remained faithful to the book's global orientation: the project of a general poetics that would be a poetics of genres. Yet, contrary to the German title, the French title no longer indicates that it is Dichtung in the etymological sense - that is, fiction as poiesis - which stands at the center of this theory of genres. Indeed, Hamburger's generic theory rests upon an analysis of what she thinks constitutes the irreducible specificity of literature: its fictional character. Her efforts are directed at "a theory of language, which investigates whether that language which produces the forms of literature [formen der Dichtung] . . . differs functionally from the language of thought and communication, and, if so, to what extent" (3). Now, these differences are clearer in the case of fiction. Tending thus to identify literature with fiction, she aims to replace the traditional definitions of literature - that is, aesthetic conceptions - with a logico-linguistic definition based upon traits internal to the literary utterance. At the same time she challenges the pragmatic and contextual criteria for fiction advanced most notably by Anglo-American philosophy and implicitly accepted by narratology.

To the extent that Hamburger defines literature by its specific place in the general system of language, her poetics is largely dependent upon a general linguistic theory. This theory is neither syntactic, nor semantic, nor pragmatic, but gnoseologic. In effect it postulates that, in its deep structure, language is an utterance-system [systeme d'enonciation] definable as a relation between an utterance-subject - which, "being fixed in language, is therefore the . . . analogue to the cognitive subject or subject of consciousness" (39) - and an utterance-object, which is that upon which the utterance bears. This subject/object relation is not a fact of communication, but is inscribed in the very interior of the structure of language: the utterance-subject is thus distinct from the I-Producer of communication, which is purely discursive or pragmatic. Hence every utterance, whatever its mode of utterance, is the utterance of an utterance-subject, apropos an utterance-object: for Hamburger, a question, a command, a wish are all equally as much utterances apropos some object as an assertoric proposition is. The central thesis of this conception of language resides in the affirmation that "all statement is reality statement" (33). Since the utterance is defined not by its object but by its subject, this thesis does not mean that every utterance bears upon some actually existing object, but that it is produced by a real person, an utterer: "statement is constituted only through a genuine, real statement-subject" (45), that is, by a real I-Origo about which we can raise the question of its place in time (even if in certain cases the response is that this place is not important - as when we are faced with a theoretical utterance-subject such as is embodied for instance by a mathematical theorem). It is obviously this thesis that makes possible the construction of a polar opposition between fiction (which does not provide for an actual utterance-subject), and the set of actual utterances [enonces de realite, "reality statements"],(1) which is to say in fact the utterance-system as such.

So it is to language as a system of actual utterances ["reality statements"] that Hamburger comes to oppose the system of literary genres. On this basis she will take account solely of discursive practices diverging in an observable way from the linguistic utterance-system: all genres, such as autobiography, history, the essay, etc., that are in fact actual utterances will be excluded from her system of literature.

2

The genre that diverges the most from the utterance-system, the only one in fact that it is really possible to oppose to the system in terms of its particular logic,(2) is fiction. Under this term, Hamburger groups four forms of utterance: third-person fictional narrative, drama, the lyrical narrative ballad, and cinematic narrative. It is upon third-person fictional narrative that her most detailed analyses are focused, because it alone, she thinks, manifests the structure of fiction in all its potentialities.

Hamburger begins by presenting a certain number of linguistic phenomena that are found in third-person fictional narratives and that she thinks impossible in an actual utterance. It is this analysis that undoubtedly constitutes the most celebrated aspect of the book and the one that has caused the most ink to spill since its appearance in 1957. I will limit myself to discussing briefly the phenomena that are concerned.

First of all it involves the use of verbs describing interior processes (for instance, thinking, reflecting, believing, feeling, hoping, etc.) applied to a third party. But, according to Hamburger, outside of fiction these verbs can be applied solely to the first person: they are only utterable on the basis of the subjective interiority they describe, since we only have access to our own interiority. 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" (83).

The second phenomenon is in some way an extension of the first: it involves techniques of free indirect discourse and interior monologue. By different means these lead to the same result, namely the direct representation of a third person's interior life. Here too discursive forms are involved that are impossible in a nonfictional narrative in the third person, and this for the same reason as the one opposed to the use of verbs of feeling: in an actual utterance a third person can only be viewed as an object and not starting from his subjective interiority.

A third trait consists in the combination of situational verbs (for instance getting up, going, sitting down, passing a restless night, etc.) with utterances concerning events very remote in time, or those of indeterminate date. Only in fiction can we use a statement like this: "Toward the end of the 1820s, when the city of Zurich was still surrounded with extensive fortifications, on a bright summer morning in the middle of the city there arose from his bed a young man . . ." (Gottfried Keller, Zuricher Novellen, qtd. 94). In an actual utterance "the continuity between a young man's getting up out of bed and the assertion that the city of Zurich, where this took place toward the close of the 1820s, was surrounded with extensive fortifications, would not be possible" (95).

A fourth difference of fiction lies in its use of dialogue. This, if not impossible, is at least displaced in actual utterances, unless the words that it reproduces are situated in a past very close to the time of utterance: "Like the narrated monologue, dialogue, too, has as its indigenous locus only third-person narration, i.e., pure fiction. For only here can narration so fluctuate that 'reporting' and dialogue blend within the unity of the narrative function" (176-77). It is in the name of this thesis, a surprising one as it must be admitted, that Hamburger will elsewhere come also to criticize the use of dialogue in first-person narratives (which, we will see, are situated outside the logical field of fiction).

A fifth trait: the use of spatial deictics along with the third person, and the combination of temporal deictics with the preterite and the pluperfect. In a fictional narrative in the third person, spatial deictics no longer fulfill any "existential function" (132); they do not organize the actual spatiality of the author's and reader's field, but are connected to fictional characters, and this, according to Hamburger, signals that they cease to function as deixis to become "symbol[s]." They no longer allow any effective orientation (since they are situated in a fictional space, and not in real or pretended space) and are in some way just lexemes connoting spatiality as such. Temporal deictics violate the rules of the system of actual utterances in an even more visible way: it is only in fiction that it is possible to combine terms like 'today' and 'tomorrow' with the preterite, or 'yesterday' with the pluperfect. This shows that the temporal axis of the narrative is not that of the author or of the reader, and hence is not the axis of the system of actual utterances but that of the characters' fictive existence.

Finally, a sixth and last trait: the atemporal function of the preterite. This trait is in fact a simple consequence of the other phenomena already enumerated. In fictional narrative the preterite no longer indicates the pastness of events as they are reported at the moment of utterance: the events recounted are not recounted as past but as unfolding themselves in the "Here and Now" [hic et nunc] of the fictional characters. This does not mean that they unfold in the present, but, more radically, that they function outside of the temporal system as such, in the (atemporal) fictive present of the characters. The "grammatical index" [objektive grammatische in the German ed. (70)] of this change of function Hamburger finds in the combination of the preterite (and the pluperfect) with temporal deictics, a combination that would be impossible if the preterite functioned as a past tense. Another confirmation (a gnoseological one(3)) is found in the use that the fictional narrative makes of verbs describing interior processes:

when we draw upon our logico-psychological experience of ourselves and recall that we can never say of another real person: he thought or is thinking, felt or is feeling, believed or believes, etc., we realize that when these verbs appear in a narrative the preterite in which the story is narrated becomes a meaningless form, if one takes it to be the tense which designates the past. (82)

It follows that the technique of free indirect discourse points a fortiori in the same direction.

From these analyses, some of which, it must be said right away, are debatable, Hamburger draws a radical conclusion concerning the status of fiction (that is, within the domain of the verbal arts, the status of "epic" narrative in the third person, of drama, and of that "special" [i.e., mixed] form which is the lyrical ballad): fictional discourse is not an act of utterance. As an act of utterance, it should in fact be uttered by a subject about an object. Now, on the one hand, we have seen that there is no utterance-subject as to which the question of his or her position in time can legitimately be raised: the question cannot be raised in fictional narrative, which is situated outside of the temporal system. On the other hand, in an act of utterance a third party can only be the object of the discourse: but fictional characters in third-person narration are represented in their interiority, as subjects, as fictional I-Origines. From this it follows that fiction is not based on a real act of utterance, and thus that it is not an utterance. But - and it is here that Hamburger turns her back on pragmatic or communicational definitions of fiction - neither is it a pretended act of utterance; for this it would have to have a pretended utterance-subject, one, that is, imitating a real utterance-subject, whereas in fact the six phenomena just reviewed show that it does not have an utterance-subject at all. Fiction does not imitate any utterance, because it escapes the utterance-system as such.

The disappearance of a real I-Origo, and hence of an utterance subject, is correlative to the emergence of the fictional I-Origines that are the characters: they can become the center out of which the narration radiates, because the real I-Origo has disappeared. There is moreover a causal relation between the emergence of fictional I-Origines and the functional change of the preterite: it is because the fictional narrative is anchored in the Here and Now of fictional characters that the preterite loses its normal function, which is that of recounting past events (real or pretended). Otherwise put: if fiction does not recount past events, it is because it creates fictional characters acting Here and Now, hence because it leaves the field of real (or pretended) utterances and the temporal system tied to a real (or pretended) speaker. One important consequence that flows from this is that, in fiction, "narration itself is not what is of principal importance" (179): as there is no longer an utterance-subject, the act of telling can no longer be distinguished from what is told. In fictional narrative the subject/object structure of reality statements is eliminated: "epic" fiction is not the narrative utterance of a subject concerning an object, which would be the event recounted. Events and their narration are one and the same thing: the fictional world does not maintain a relational tie (a subject-object relationship) with the narration, but solely a functional tie. Fictional narration is not the act of a (real or pretended) subject, but a self-engendering, fluctuating mimetico-poetic function, since it is embodied as much in the narrative (in the narrow sense of the term), as in the dialogues or generalizations, without which it would be possible to distinguish functionally between these different elements. Fiction is thus a narrative without a narrator, a narrative that tells itself. For if it has no fictional narrator, no more is it the author who tells it: the latter stands in relation to the text as its creator and thus remains exterior to the fictional world (he is not its utterance-subject) in the same way that the painter remains exterior to the picture. The parallel could be extended: a narrative is no more the author's telling than a painting is the painter's field of vision.

Hamburger asserts that this conception of fiction merely develops the Aristotelian conception of mimesis as a presentation of men in action. Thus "epic" narrative is not the mimesis of a speech act (a ludic pretense of a serious narrative), but a mimesis of reality: it is an appearance, an illusion of life (Schein des Lebens).

We see that the definition of fiction as a "semblance of life," the refusal to consider it a pretended speech act, and the thesis of the disappearance of any enunciative action (and of any narrator) all go together. It is thus that Hamburger can say that narrated fictional life "is just as independent of an utterance-subject as is 'real' reality" (136). It is in this same perspective that her numberless comparisons between the author and the painter should be read: the author handles the narrative function as the painter handles a brush and colors. In fiction

the process of "making" is at work, in the sense of forming, shaping, fashioning; here is the workshop where the poietes or mimetes creates his [characters], using language as his instrument and construction material just as the painter uses color and the sculptor stone. Here literature is completely within the domain of the plastic arts, which create the semblance of reality. (233)

In fiction, narrative is transmuted into a plastic presentation: we move from the domain of language to that of appearance, of Schein. It is no doubt within this conception that the most intimate kernel of Hamburger's theory is located - but also its rather mystical point.

According to Hamburger, everything that she has said about third-person narrative goes as well, with some restrictions, for drama, because the two poetic forms constitute a single genre, that of fiction: like narrative, drama is a mimesis of life, a creation of fictional characters acting Here and Now. The restrictions are connected with the fact that drama is only a subgenre, "a sort of part extracted from the epic substance" (qtd. from the German ed., 175: gleichsam herausgeschnitten): since the dialogue is its only means of presentation, it cannot give us the direct mimesis of interior life, which narration can accomplish by means of variations of the narrative function, that is, thanks to free indirect discourse, interior monologue, or the use of verbs of feeling.(4) This postulated identity of "epic" poetry and drama obviously entails that we cease to see in "this difference of presentational form, i.e., narration and the formulation of character through dialogue, . . . a difference in genre" (196). Thus the fact of being represented is not an essential characteristic of drama, but is merely a consequence of its specific mimetic construction, namely its dialogic mise en forme: "For the locus of drama within the logical system of literature is determined solely by the absence of the narrative function, by the structural given that the [characters] are formed through dialogue" (198). "But the stage is the non-literary partial function which the verbal art work can (not must) make use of" (215). The refusal to consider effective material representation as an internal characteristic of drama finds its correlate in the refusal to treat "epic" poetry as a narrative situation: it is by reducing the two forms to their presentational content, to the appearance of life that they realize through language, that it is possible for Hamburger to unite them in a single genre.

3

Third-person narrative and drama (as well as cinematographic fiction and the mixed form of the ballad - which I disregard here in order not to overextend this summary) delimit the domain of the fictional genre. Furthermore, according to the theory of language which Hamburger defends, the system of language allows for only two sorts of logical status, that of actual utterances and that of "fictive" utterances. These two theses, taken together, obviously make it difficult not only to characterize other genres that she recognizes, namely pretended narrative in the first person and lyric poetry, but also to analyze their correlation with fiction in the global framework of a generic system.

Let us begin with first-person narrative. It is situated outside the system of fiction without, for all that, being an actual utterance. It is not an actual utterance because it has no "real" utterance-subject. But no more so is it a self-engendering mimesis: it indeed possesses an utterance-subject who is a pretended subject. This implies that every first-person narrative presents itself as an actual utterance, appears as a "historical document" (312). It is thus subject to the structural laws of the actual utterance, and notably to the subject-object polarity: "the first-person narrator does not 'engender' that which he narrates, but narrates about it in the same manner as in every actual utterance: as about something which is the object of his statement, and which he can only present as an object" (317). First-person narration is not a mimesis of reality but a speech act: it is a "reigned reality statement" (313). But if it is subject to the structure of the actual utterance, this means that the forms of fictionalizing representation are forbidden to it: no verbs of feeling applied to the third person, no free indirect discourse, no monologue, and no dialogue, at least when these are situated in a receding past (for instance, in fictional memoirs). Hamburger goes even further, since she adds that these techniques of fictionalization can have no place at all in a first-person novel, not only insofar as they apply to the third person, but "even where the I-Narrator is himself the case" [German ed., 275]. Of course, in varying degrees most first-person narratives violate these rules (especially those concerning dialogue and the absence of verbs of feeling related to the I-Narrator): this fact is nothing other than an index of partial fictionalization, evidence of the bastard status of a good number of first-person narratives.

Contrarily to fiction, the reigned first-person narrative does not find itself in a categorical opposition to the system of actual utterances: the distinction between a veridical narrative in the first person (for instance an autobiography) and a reigned narrative that closely imitates serious speech acts can only be drawn from context, that is, from exterior knowledge and not from internal traits. Thus first-person narratives do not form a true part of the system of literature [Dichtung]; it is not a canonical genre, but simply a "special [i.e., mixed] form."

If first-person narrative is outside the "system of literature," the status of lyric poetry is more complex. Repeatedly, Hamburger opposes the lyric pole to the epic-dramatic in affirming that together they encompass the entire space of the system of Dichtung. Thus Dichtung will be constituted by fiction and lyric poetry: this explains why the latter is discussed immediately after the epic-dramatic genre as an entirely separate genre and is not included in the "special forms." Hence lyric poetry should also find itself in a categorical opposition to actual utterances. Yet lyric has nothing to do with fiction, and Hamburger insists on the fact that we situate lyric poetry "on an entirely different level of our imaginative life" (5): it creates no acting persons, fictive I-Origines, but it is the utterance of a subject, the "lyric I." Since the system admits of only two terms (if we except the bastard pole of the reigned actual utterance), there is no choice but to maintain that lyric poetry produces true actual utterances, so that the "lyric I" is a real utterance-subject: "The much-disputed lyric I is a statement-subject" (234). But if this is the case, how can it still form a system with fiction?

Hamburger's attitude in the face of this dilemma, into which she has been led by her wish to posit a system of poetry, is ambiguous, and the section devoted to lyric poetry is incontestably the weakest (and least innovative) in the book. In it one can no doubt discern the result of a residuum of piety for the epic-dramatic-lyric triad, which leads to the implicit presupposition that the three terms could be incorporated into a single categorical unity (the system of Dichtung). All sorts of incoherencies follow. She begins by asserting that the single criterion which allows of a distinction between lyric poetry and other actual utterances is of a contextual order. The lyric form itself therefore is not determining: a sacred song by Novalis belongs to lyric poetry (and thus possesses a lyric utterance-subject) when it placed in a collection of poetry; on the contrary, when it is placed in a collection of prayers, it is no longer the utterance of a lyric I and thus ceases to be lyric poetry. Thus, "the lyrical genre becomes constituted through the so to speak 'announced' intention of the utterance-subject to posit itself as a lyric I, and this means by the context in which we encounter the poem" (241). If this is the case, it is hard to see how lyric poetry could form a system with fiction, which is categorically distinguished from actual utterance(5): according to this first definition, lyric poetry will rather tend in the direction of "special form."

However, Hamburger's later analyses all aim to establish, on the contrary, a structural criterion, allowing the opposition, in the interior of the field of actual utterances, of two sorts of utterances: the communicative utterance and the lyric utterance. Thus the communicative utterance is directed toward the object-pole, and lyric poetry is characterized, on the contrary, by the fact that utterances retreat from this pole: they "are not oriented toward an object-nexus, nor are they controlled and directed by it. They do not form any objective context, any context for communication; rather, they form something different, something which we shall call a sense-nexus [Sinnzusammenhang]" (249). The utterances are no longer in a vertical relation with the object-pole, but are structured horizontally by connection with each other and in relation to the lyric subject which they express: "the lyrical statement-subject does not render an object of experience, but the experience of the object" (278). In lyric poetry, the object has no pertinence except as an object subjectively tested, except as an object situated within the subject. Does this amount to a truly structural demarcation? We might try to raise it to a contextual criterion by postulating that it is not the opposition of two forms of utterance that are involved, but rather of two conditions [etats] of utterance - the communicative condition and the lyric condition. But this attempt would force us to consider as lyrical every utterance shorn of its communicative dimension, which seems to be an overly liberal proposition at best. Therefore let us admit that a structural criterion is involved, which is to say a criterion that has an effective, observable base. In this case, lyric poetry could truly form a system with fiction, since it would be categorically distinct from the communicative actual utterance. The "lyric subject," an indeterminate subject, would thus be the receptive hypostasis of the structural locus of the utterance-subject, left vacant by the withdrawal of the communicative subject. Yet this solution also poses its problems, this time at the level of general linguistic theory. In place of two categorical states [statuts] of language, three must be recognized: the communicative actual utterance, the noncommunicative actual utterance, and the fictional utterance.

Within this systematic structure there is evidently no place for the reigned first-person narrative, since the latter does not fit into a categorical opposition to the system of actual utterances. In order to do this it would be necessary to put the question of systematicity in parentheses, and thus that of the distinction between contextual determination and structural determination. In this perspective, fiction in the first person could figure as a literary form in the full sense of the term, occupying an intermediate place between lyric poetry and fiction, that is, between noncommunicative actual utterance and fictional utterance.

4

This brief summary gives only a very imperfect account of a book that Gerard Genette rightly emphasizes is, "in the details and steps of its analyses, terribly convincing" (Preface xviii). The core of the book, the aspect of it that is simultaneously the most innovative and the most provocative, obviously consists in the theory of mimetic fiction (and the related question of the status of first-person narrative). It is to this problem that I want to devote some brief remarks (I will leave aside everything relating to dramatic fiction).

First it is necessary to see clearly what it is that Hamburger's theory involves us in. I have already drawn attention to the multiple comparisons the author draws between fiction and plastic art: thus, in a quasi-Heideggerian passage we learn that fiction possesses a "presentifying quality"

which is none other than that conveyed to us by a painting or a statue, a quality of being-there (Da-sein), of being-eternal [Immersein], a "static Here and Now," which is the basic meaning of the German concept of Gegenwart as well as of the [Latin] concept representare, with respect to which the temporal meaning is secondary and derived. The Now is contingent upon the Here, etre present, not the Here upon the Now. (97-98)

The representative character of fiction abolishes, at least tendentially, its status as a speech act: the fictive world is certainly linguistically presented, but its presentational act is abolished in its content. Language is the material of "epic" poetry, not its communicational channel: the "semblance of life" is presented in and by the work, not constructed by the reader beginning from the information that will have been transmitted to him by the narrator. Ultimately, fiction is not a narrative that we read, but a picture upon which we gaze.

The negation of the narrator, which constitutes one of the most provocative traits of Hamburger's theory, is thus in fact only a consequence of her more general refusal to consider mimetic fiction as a speech act. Moreover, what goes for the question of the narrator goes also for other aspects studied by narrative analysis. This is the case with temporality: if narrative is essentially the representation of characters acting Here and Now, in an atemporal present, questions of order and duration (see Genette, Narrative Discouse 33-112) concerning the relationships between the time of the narrative and the time of the story (for example: what is the order of the narrative in relation to the chronological order of the events reported?) no longer have any place. Indeed, if we follow Hamburger's theory, the narrative (the narration of events) would not be relatable to the story (the events narrated), very simply because the latter has no relevance in fictional narrative (thus "the question as to what the 'point' at hand is in a novel elude[s] clear-cut, fixed assignment" [167]). The same goes for phenomena that Genette has grouped under the headings of mood and voice (Narrative Discourse 161-262). Hamburger maintains for example that the distinctions between narrative of words and narrative of events (distinctions which in Genette belong to questions of mood) have no functional status. Narrative (Bericht) and dialogue are not opposed as a diegesis of events and a mimesis of words, but constitute, quite simply, two variables of the same narrative function, qualified as "fluctuating" [cf. 176]. The problem of localization (that is, the question of knowing from what viewpoint events are narrated) likewise has no reason for being: to the extent that narrative fiction is fluctuating, at the interior of the logical domain circumscribed by characters such as fictive I-Origines, changes in localization will only be surface phenomena, without importance at the level of the fictional world, because, in virtue of an ontological necessity, this fictional world is always presented in the mode of internal localization. Hence this general conclusion: "relations in fiction escape all assignability" (167).

One sees the interplay of these questions: acceptance of the theory of fiction proposed by Hamburger entails surrendering entire sectors of narratology as being, at best, superfluous. I do not think that one can escape the problem by postulating that narratology is interested in "surface structures" (narrative) whereas Hamburger has interested herself in "deep structures" (fiction).(6) Indeed, according to this distinction the narrator would only have a function at the level of surface structures. Yet the whole of Hamburger's theory rests on the distinction between narrative with a pretended narrator and narrative without a narrator, which is to say that this distinction occurs very much at the level of (what she thinks to be) deep structure. According to Hamburgerian logic, to speak of a narrator in a third-person narrative is not to speak at the level of surface structures - it is to commit a category mistake. The act of situating the question of the narrator at the level of some kind of surface structure thus amounts in fact to denying that the distinction between utterance and narrative without utterance has a categorical status, which is to say that this amounts to opposing Hamburger's theory. Thus, in the matter of fictional narrative in the third person, her theory seems to me to be really incompatible with narratology, or at least with certain aspects of it.

At first sight one may think that the definition of fiction as a mimesis of life accords with the deliverances of common sense. Because everyone is in agreement that fiction, in presenting characters in action, results in the creation of a fictive life, an apparent life. Indeed, we readily speak of fictional characters, their adventures, their circumstances, etc., as if we were speaking about real people. In reality however Hamburger's theory is much too specific to be able to agree with common sense. When we say that fiction creates a semblance of life, we are in fact drawing no distinction between third-person narrative and first-person. But according to Hamburger, this distinction is crucial. In the second place, if common sense tends to credit drama with a mimetic effect superior to that of narrative, is this not the case because the former makes the characters take flesh in our tridimensional reality? Hamburger maintains the contrary, since drama, being limited to dialogue, cannot directly represent the characters' interior life. Finally, according to our spontaneous conceptions, the mimetic effect is not limited to fictional literature: a historical narrative can produce exactly the same result. All this seems to me to demonstrate that, for common sense (it must be excused if I make too much of this), 'mimesis of life,' rather than being the logically defining characteristic of a certain class of narratives, is an imaginative effect induced by any narrative at all in the reader or auditor. This means that at the level of common sense there is not necessarily a contradiction between saying that fiction is a 'mimesis of life' and saying that it is the utterance of a narrator, hence that it is (eventually) the mimesis of a speech act. In Hamburger on the contrary the two phenomena exclude each other.

Quite clearly, the author of The Logic of Literature doubtless never meant to satisfy common sense (even if a part of the persuasive force of her formulations comes from analogous formulations of common sense). The fact that she distances herself from it is thus certainly not a decisive objection to her theory: it simply shows that we cannot appeal to the deliverances of common sense to affirm the theory's correctness.

Thus the question of its plausibility poses itself at the level of its status as a technical theory. To test the degree of plausibility of such an item, we have three rather reliable means available. The first is to see which generally recognized phenomena (recognized in other technical discourses, i.e.) are denied by the theory in question, and which phenomena by contrast it incorporates better than such and such a concurrent theory. A second is to see whether the theory effectively explains everything that it claims to explain. And the third is to compare the theory with concurrent explanations in respect of what could be called their 'conceptually economical character.'

Let us begin from the first point. Hamburger's theory rejects as nonpertinent a large part of the phenomena analyzed by narratology. As I have already discussed this problem, I will simply stress one point: these phenomena have a textual presence of such pregnancy that the fact of being unable to integrate them hardly speaks in favor of a theory. But could it not be that in compensation the phenomena which Hamburger explains pose problems for narratology? For the latter indeed fiction in the third person is a pretended (serious) narrative, hence the mimesis of a speech act (as first-person narrative is). Hamburger's analysis claims to have shown that certain phenomena are practically never encountered in (serious) narrative in the third person, while they are omnipresent in fictional narrative. How can it be maintained that narrative in the third person is the pretense of a serious narrative, if some of its most important characteristics are impossible in the latter?

Let us note first that the phenomena Hamburger enumerates are far from being as evident as she claims. Thus some of them, contrary to what she maintains, are perfectly represented in serious, third-person narrative, whether written or oral: this is the case of dialogues, verbs of situation, and verbs of feeling applied to third parties. Sometimes her interpretation seems risky at best: this is notably the case in respect of her affirmation of the atemporal character of the preterite. Two counterexamples will suffice to show this. The first is that of iterative narrative: if the preterite had no temporal function, every iterative narrative, that is, every "single narrative utterance [that] takes upon itself several occurrences together" (Genette, Narrative Discourse 116), would be impossible, since every recapitulation presupposes the logical relevance of the category of the past tense in relation to the moment of narration. The other is that of prolepsis, that is, of the anticipation of future events. Even if it is less widespread in third-person narrative than in narrative in the first person, it is nevertheless not absent. Now, when we are confronted by a sentence of the type, 'Later on he came to regret this decision,' we cannot fail to realize that the act of utterance is posterior to the decision of which we are told, and that this act of utterance is referable to a narrator who is situated outside of the Here and Now of fictional characters.

In the second place, 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 localization 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 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.

In reality, the prohibition is not coextensive with serious narrative as such: thus Freud's case histories conform in large part to the principle of heterodiegetic narrative with internal localization, since Freud, by virtue of his psychoanalytic knowledge (whether real or imagined does not matter), does not hesitate to describe the inner feelings experienced by his patients in their early infancy. The true distinction in principle between serious and fictional narrative consists in the fact that in the case of serious narrative we require the author to explain to us the source of his acquaintance with the interior life of his hero (or character), whereas in the case of fictional narrative this question is excluded by the rules of the game. Heterodiegetic narrative with internal localization is therefore hardly a logical impossibility for serious narrative; its utilization very simply presupposes, as does every serious narrative (whether bearing upon interior or exterior events), that the author be prepared to legitimate the provenance of his information.

The distinction established by Hamburger is therefore not a logical distinction, and fictional heterodiegetic narrative with internal localization does not come into a categorical opposition to serious narrative. The fact that its practice is much more widespread in the domain of fiction than in the domain of serious narrative can be explained quite simply by the fact that it is much more easily achievable in the first case, since the question of the provenance of an acquaintance with the intimate life of one of the characters cannot arise. And even if it is true that the techniques of representation for inner life are more developed and more numerous (free indirect discourse, or interior monologue for example) in fictional narrative with internal focalization than in the corresponding serious narrative, this difference nevertheless does not at all justify the absolute opposition that Hamburger erects between narrative with a narrator and narrative without a narrator.

Let us turn to the second question: does her theory actually explain all the phenomena that it claims to? To the extent that she affirms that there is an absolute, categorical boundary between fiction and nonfiction, the question can be reformulated in this way: is her definition of fiction sufficient and necessary, that is, does it enable us to discriminate between a fictional third-person narrative and a third-person narrative that is not fictional, in every case?(7) Her definition appears sufficient, at least insofar as free indirect discourse, monologue, and the use of temporal deictics tied to the temporality of fictional characters are concerned: whenever we come across these phenomena, we can be reasonably sure of finding ourselves in a fictional narrative. But it is easy to show that her definition elicits no necessary criteria. Let us take the case of an invented third-person narrative, which abstains from free indirect discourse, interior monologues, and temporal deictics connected to the preterite and the pluperfect - that is, from those among the phenomena which Hamburger has isolated that could reasonably be accepted as sufficient criteria of fictionality. We have no need to require in addition that it abstain from dialogues to report discourse, nor situational verbs applied to temporally indeterminate events, since these phenomena are found too extensively in actual utterances to be able to function as sufficient criteria for fictionality. As for the atemporal function of the preterite, it no longer needs to be taken into account: being just a consequence of the other phenomena, it disappears along with them. Thus we would have a nonveridical, third-person narrative that would comply with all the internal constraints of actual utterance and that would only be distinguished from them by nonacceptance of an external presupposition: it relates the story not of effectively existent characters, but of invented characters. Let us furthermore suppose that we have - coming from the paratext, for instance - a generic indication, so that the reader may know that this is an invented narrative. Narratives of this kind [genre], or those approaching this model asymptotically, clearly exist, since what comes under this description is quite simply heterodiegetic narrative with strictly external focalization, for instance Hemingway's "Hills like White Elephants" In this case the narrative is of course not a serious pretense of a real third-person utterance, and so is not a deceptive narrative (which would be a lie, thus belonging to the system of actual utterances), since we have presupposed that the reader knows that it is an invented narrative(8): we find ourselves very much facing a narrative situated outside the system of actual utterances. According to narratology this consists quite simply of the playful pretense of a serous speech act, in the guise of a truthful narrative. Hamburger's theory, by contrast, does not allow for the possibility of such a narrative, since it posits only two forms of third-person narrative: serious narrative and fictional narrative, the model for which the theory takes from the phenomena of fictionality.

Let us turn to the third question, concerning the reciprocal advantages of Hamburger's theory and the narratological approach. We might begin from the fact (admitted by Hamburger) that first-person narrative continually violates the rules that give it its status as a pretended homodiegetic actual utterance, having recourse to techniques of fictionalization. Once we look further than the historical field that the author of The Logic of Literature privileges (namely that of nineteenth- and twentieth-century narrative), we discover that, in third-person narrative, the phenomena of fictionalization are far from being so omnipresent as her theory leaves it understood to be. Then why not very simply recognize a parallel between narrative in the first person and the third person? Put another way, just as homodiegetic narrative with internal/external localization (internal with respect to the narrator, external with respect to the other characters) would be the unmarked state of first-person fictional narrative, heterodiegetic narrative with external focalization would be the unmarked state of third-person fictional narrative. Thus, instead of seeing characteristics of the essence of third-person narrative in the phenomena of fictionalization, one can see it in terms of deviations from the unmarked state of such narrative.

However, the situation is not so simple. On the one hand, the deviation of third-person fictional narrative from serious narrative is larger than the deviation of first-person narrative from its corresponding serious models. Essentially this involves the narrative techniques that are connected to the description of the mental life of the third-person characters. On the other hand, whenever homodiegetic narrative reports the intimate thoughts of a third party in detail and in the form of a direct transcription, we automatically perceive this as the violation of a rule. It is not at all the same with third-person narrative, which very rarely realizes a strict "zero degree," and which, in the majority of cases, represents, at one time or another, third parties viewed from the interior, without this appearing incongruous to us. Thus, the deviation from the unmarked state is not experienced in at all the same way in the two cases. Now, it is exactly this difference in the reader's fundamental attitude toward these two sorts of narrative that Hamburger's theory allows us to explain. In contrast, for narratology this seems to pose an insurmountable problem: if every fictional narrative is a pretended speech act, how is it that in the case of first-person narrative we experience deviations from the imitated speech act as violations, while this is not at all the case in third-person narratives?

It seems to me that it is possible to extricate narratology from this unfortunate situation, but for this the notions of 'narrator' and 'pretended discourse' have to be made precise. In regard to 'narrator,' there is sometimes too strong a tendency to reify this into a persona, and even into a character. In reality, according to the communicational conception of narrative (which implicitly lies at the foundation of narratology), it should simply be considered as a pragmatic presupposition of any narrative: until the contrary can be shown, there is no imagining a narrative without a narrator; where there is narrative, there just has to be someone who tells it. Against this, no fixed role corresponds to such a presupposition: the narrator can be situated anywhere along the axis which runs from the fictional character at one end to the simple, unspecified author-narrator at the other. Furthermore, in most cases the narrator remains entirely indeterminate, except insofar as he is the utterer of what is uttered: this is by definition the case of every anonymous narrative without a fictively constituted narrator (for instance popular tales). Hence the fact that in heterodiegetic narratives with external focalization no fictional narrator emerges as a persona backing the narrative is not an argument against the idea of the narrator: the question of his communicational 'existence' ought not to be confused with that of his status as a character.

This brings me to the notion of a 'pretended speech act.' It is clear that the term actually designates two different phenomena: the author can pretend to tell a true story, that is, he can pretend to make referential assertions (then we would have a pretended speech act); but he can also pretend that it is someone else who makes referential assertions (then we would have the pretense of a speech act).(9) In the second case there is a fictional narrator, while in the first case the narrator's status is blended with that of the author (which does not mean that they are existentially identical, but very simply that we possess no means for distinguishing them). Hence differences in the narratee: the narratee of a nonpretended narrator is always the reader, while that of the pretended narrator is much less often so (for instance a fictional journal is not addressed to the actual reader). Put differently, the communicational status of the narrator is not the same in the two cases: outside of the pretense in the first case, it is part of the pretense in the second.

It is important to notice that this distinction cannot be superimposed upon the distinction between third-person narrative and first-person narrative. A third-person narrative may very well be told by a fictional narrator: it is sufficient that the indices of the narratorial situation turn out to be irreconcilable with what we know (or think we know) about the author. Therefore, the pretended author of a fictional historical chronicle, the pretended author of a fictional biography, etc., are definitely fictional heterodiegetic narrators. It is undeniable that, in numerous cases, we make decisions simply according to the principle of economy: so long as indices of the narrator's fictionalization are lacking, we identify him without further examination with the author. From the standpoint of the logic of actual utterance, this step will remain dubious, since the supreme (serious) pretense is the one that leaves no traces. But the question of the narrator's status is quite simply a problem of communicational economy: as long as nothing forces us to postulate a dual uncoupling with regard to actual utterances (the absence of illocutionary force and the pretense of an utterer), we fall back upon the real context (the author) and accept that it is the narrative that is pretense (rather than the utterer).

Where there is a fictional narrator, one conceives of him as situated within the pretense, as a part of the fictional universe. This implies that he is subject to the logic of plausibility, that is, to the rule which maintains that we have the right to treat all characters' actions as going on in a 'possible world,' and thus as things that could be happening or could have happened really. The fact that this rule is also applied to the writer-narrator explains why, when the narrator in (for instance) a fictional autobiography violates the rules of plausibility for serious narrative by representing the interior life of third-party characters directly, it generally shocks us. Indeed, in a serious narrative, the representation of the interior life of a third party is only possible if it is legitimated by an indication of the source from which the information comes. Hence, when this legitimation proves to be lacking in an utterance that belongs to the pretended universe, we sense a break in plausibility. This fact also explains why homodiegetic narrative departs less from its serious models than heterodiegetic narrative: the model imitated, the speech act imitated, belongs here to the fictional universe, that is, it is subject to the rules of plausibility.

In contrast, when the narrator's status is left indeterminate (as is the case in many heterodiegetic narratives), its utterance [acte d'enonciation] tends to be referred to the author, and thus to be excluded from the field on which the pretense bears; here then it is the pretended narrative which is subject to plausibility, but not the narrative's narration, which is accepted as a given absolute. At the level of the institution of pretense we do not apply the rules which count for the pretended universe. Or to put this another way: in the case of narrative with a pretended author, the question of the acquisition of the information which sustains the narrative belongs to the same field as the fiction; in the case of narrative with an indeterminate narrator, in contrast, the question is excluded from the field of fiction. In the second case we implicitly identify the narrative act with the act that gives birth to the pretense, while in the first case the narrative act belongs to the pretended world.

It seems to me therefore that Hamburger's theory possesses a real, irreducible validity: the opposition that she has established between "fiction" and "pretense," even if she has put the boundary between them in the wrong place and has overinterpreted their differences, which actually correspond to a distinction in the communicational status of the narrative act, depending upon whether it is a pretended speech act or the pretense of a speech act. In this sense her book is a major (although unintended) contribution to narratology. Hamburger has put her finger on a phenomenon that the latter no doubt has a tendency to underrate (perhaps because, sometimes,(10) it has wanted to isolate the 'text' at any price from the author), namely the fact that the narrator of a fictional narrative oscillates between two poles, each having a different communicational status, depending on whether it is pretended or remains, on the contrary, outside the pretense. The distinction is important, because, as Hamburger's analyses have shown, it has effects at the very level of narration. At the same time it illustrates perfectly the 'impure' status of fiction, which, as a nonserious speech act, is free to emphasize its invented aspect or on the contrary its pretended aspect, just as the narrator can oscillate between the author at one pole and a pretended character at the other. But this is not because the pretended narrator is unassimilable to the narrator who establishes the pretense, or because the narrative act possesses a different signification in the two cases, so that one must take it that the nonpretended narrator is not a narrator at all and that the narrative tells itself; when no index of fictionalization appears, the most economical course still consists in taking it that the author is telling the story. The fact that the narrative may be a fiction (hence, that it is told by the author) is not opposed to its being a pretense (hence, that it is told).

Translated by David Gorman

TRANSLATOR'S NOTE: This study originally appeared in 1987 under the title "Fiction, feinte et narration" in Critique ([#]481-82: 555-76). It discusses the major work of Kate Hamburger (1896-1992), Logik der Dichtung, which first appeared in 1957. Marilynn J. Rose's English translation, published in 1973, is based on the enlarged second edition of 1968. In the present essay, meanwhile, Schaeffer is dealing with Pierre Cadiot's French translation of the book, which appeared in 1986, and which is based upon the third German edition (1977). It should be noted that the translation which Schaeffer is analyzing is entitled Logique des genres litteraires (i.e., the Logic of Literary Genres): this orientation towards genre (which is faithful, incidentally, to Hamburger's intentions [cf. 232 of the English translation]) obviously affects Schaeffer's presentation and critique.

Triangulating between the English translation of Hamburger, Schaeffer's French text, and his citations from Cadiot's French translation of Hamburger has produced complications in places (see n, 1, for example). My general practice has been to replace the passages quoted by Schaeffer from Cadiot's translation with those from Rose's English (the same for page-citations); in places where I was unable to do so, I have indicated significant changes or insertions with brackets; passages Schaeffer quotes from the French translation for which I could find no equivalent in the English are noted with a bracketed question mark, followed by the page cited by Schaeffer from Cadiot.

Thanks to the author for permission to translate, as well as for corrections and amendation (see Notes) to this translation.

Notes

1 [Translator's note: Hamburger's term Aussage is translated into English by Rose as "statement," but into French by Cadiot as enonciation (utterance). Following my policy on translation, I have replaced enonciation in Schaeffer's citations from the French translation with "statement," even though I would have preferred "utterance," as clearer and more natural-sounding: the odd and mysterious term "reality statement," for example, which occurs on almost every page of the English translation, simply means "actual utterance." With this in mind I have rendered enonciation and various related terms as "utterance" in translating Schaeffer's words, at the expense of inconsistency with the usage in quoted passages.]

2 Beyond epic fiction in the third person and drama, Hamburger also includes lyric poetry in her system, but without managing (in my opinion at least) to show that its opposition to the system of actual utterances is situated on the same level as that of fiction. I will return to this point subsequently.

3 The German edition indicates explicitly that what is involved is an erkenntnistheoretischer Beweis (79).

4 [Translator's note: sentence modified by author for this translation.]

5 Obviously this is only true at the level of the overall structure of the narrative: not all the sentences in a fictional narrative are logically impossible utterances in the system of actual utterances, or equally well: at the level of isolated sentences the contextual criterion can intervene even in the case of epic poetry.

6 This is the solution that Franz Stanzel has proposed in A Theory of Narrative (15-16).

7 Hamburger herself states this principle: "What goes for all examplars of a genre except one does not suffice as an explanatory reason for this genre" (German ed., 287).

8 [Translator's note: clause revised by the author for the translation.]

9 See Searle. The passage which follows has benefitted from a number of

criticisms formulated by Gerard Genette in response to a previous version of this text.

10 This is not the case of Genette, who insists on the possibility of an identity between a narrator and a real author, even if he believes this to be rare (see Narrative Discourse Revisited 132-33.)

Other Works Cited

Genette, Gerard. Narrative Discourse. 1972. Trans. Jane E. Lewin. Ithaca: Cornell UP, 1980.

-----. Narrative Discourse Revisited. 1983. Trans. Jane E. Lewin. Ithaca: Cornell UP, 1988.

-----. Preface. The Logic of Literature. By Hamburger. Trans. Dorrit Cohn. vii-xix. Preface to the French trans. (1986), added to the 1993 rpt. of the English trans.

Hamburger, Kate. Die Logik der Dichtung. 3rd ed. Frankfurt/Main: Ullstein, 1977. Trans. (French) Pierre Cadiot: Logiques des genres litteraires. Paris: Seuil, 1986.

Searle, John R. "The Logical Status of Fictional Discourse." 1974-1975. Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge UP, 1979. 58-75.

Stanzel, Franz. A Theory of Narrative. 1979. Trans. Charlotte Goedsche. Cambridge: Cambridge UP, 1986.

Jean-Marie Schaeffer is directeur de recherche at the CNRS, Paris. He is the author of La naissance de la litterature: la theorie esthetique du romantisme allemand (1983), L'image precaire: du dispotif photographique (1987), Qu'est-ce qu'un genre litteraire? (1989), L'art de l'age moderne: l'esthetique et la philosophie de l'art du XVIIIe siecle nos jours (1992: translation forthcoming from Princeton UP), and Les celibataires de l'art: pour une esthetique sans mythes (1996). He is also co-author, with Oswald Ducrot, of the Nouveau dictionnaire encyclopedique des sciences du langage (1995).

David Gorman (dgorman@niu.edu) is professor of English at Northern Illinois University and the book-review editor for Style. He has published essays and other material on the history and theory of literary criticism in Diacritics, New Vico Studies, Philosophy and Literature, and Poetics Today, and is currently preparing a monograph on the work of Gerard Genette.

COPYRIGHT 1998 Northern Illinois University
COPYRIGHT 2000 Gale Group