Fabula and Sjuzhet in the Analysis of Narrative: Some American Discussions

Fabula and Sjuzhet in the Analysis of Narrative: Some American Discussions

JONATHAN CULLER

Jonathan Culler is one of the main semioticians in the Anglo-Saxon sphere. He has both adapted and criticized the European structuralist schools in his influential Structuralist Poetics ( 1975) and The Pursuit of Signs: Semiotics, Literature, Deconstruction ( 1983). He is also the author of On Deconstruction ( 1982), probably the most influential overview and assessment of deconstructive theory. Culler argues that the deconstructivist project does not delegitimize structuralist theory, but complements it. In the article reprinted here he attacks the widespread assumption among French narratologists that events at fabula level are arranged according to a true or natural order, later modified and disrupted by the requirements of narrative presentation at discourse level. Drawing on the American theoretical tradition, Culler contends that literary as well as non-literary narratives are organized around a double, contradictory logic, the logic of events, which reinforces the causal efficacy of origins and assumes the primacy of events regardless of their signification, and the order of coherence, which denies their causal efficacy and treats the events as primarily the products of meaning. Culler's thesis is that these opposed logics do not cancel each other, but rather create a tension on which the power of the narration depends, as he forcefully demonstrates in his analysis of several literary and non-literary narratives. Culler's definition of fabula as a tropological construction, the product, rather than the reality reported by discourse, is in line with Mieke Bal view in Narratology ( 1985) that the fabula is the most abstract of all narrative levels.

Although the idea of narratology and the project of a grammaire du rcit seem to have arisen and acquired their force within French structuralism and recent work in this expanding field has been dominated or provoked by the problems raised in structuralist

accounts of narrative, there has been an important American tradition in the study of narrative, whose salient moments are Henry James Prefaces, Percy Lubbock The Craft of Fiction and Wayne Booth The Rhetoric of Fiction. Inspired by these examples, critics have done much detailed and perceptive work on problems of narrative technique and in particular on narrative point of view. If one were to summarize very crudely the theoretical claims of this critical tradition, they might run somewhat as follows: every narrative has a narrator, whether or not he is explicitly identified. To interpret a narrative one must identify the implied narrator and what in the story belongs to his perspective, distinguishing between the action itself and the narrative perspective on that action, for one of the central thematic issues of every story is the relationship between the implied narrator (with his knowledge, values, etc.) and the story which he narrates.

The effect of recent French work in narratology on this tradition has been to provoke more attempts to systematize and refine the models and concepts which had often been used in an ad hoc way to interpret individual texts. So we have witnessed, for example, Gerald Prince's development of the concept of 'narratee,' as distinguished from 'implied reader' -- an important distinction which previous writers had generally overlooked -- and Seymour Chatman important work of synthesis, Story and Discourse. Indeed, narratology seems to me at the moment a flourishing area of American criticism, perhaps because it is the area in which it seems most possible to achieve integration, or at least dialogue, between an American critical tradition and various European theoretical developments.

What I am going to discuss is not, however, this tradition -- the study of point of view leavened with structuralism and formalism -but rather some recent discussions of narrative which can be aligned with this tradition in that they investigate something which the tradition of point of view studies must, as it were, take for granted. If one is to study point of view and narrative technique, if, more generally, one is to study the relationship between the discourse of a text and the story it tells, then the notion of fabula, story, plot, action -- call it what you will -- becomes the ground of one's endeavor, the point d'appui which makes the study of point of view possible. For the study of point of view to make any sense, there must be various contrasting ways of viewing and telling a given story, and this makes 'story' an invariant core, a sequence of actions which can be presented in any of various ways. Action becomes something that exists independently of narrative presentation; in principle it exists prior to any narrative presentation and could be presented in other ways. For example, when Grard Genette sets out to study the complicated temporal relationships between rcit and discours in his 'Discours du rcit,' he must assume that events of the rcit occurred in some order and that each event occurred either once or more than once ( Genette, 1972). Then he can describe the narrative presentation as a transformation of the true or original order of events. Of course, in a particular narrative it may be impossible to tell from the evidence presented whether event A preceded or followed event B, but since it is assumed that at the level of fabula there must have been a true order, this impossibility can be taken as a fact about point of view. Without the assumption that there is a true order of events prior to narrative presentation, one could not claim that the lack of order was the result of point of view.

Of course it is not unreasonable to assume that events do occur in some order and that a description of events presupposes the prior existence of those events. But in applying to the text of narrative these perfectly reasonable assumptions about the world, we isolate a level of structure, call it fabula, which we treat as something given, a constant, a sequence of events, which the narrative presupposes and which it could describe in various ways. By identifying this sequence of actions as what the text is describing, we make it possible to treat everything else in the text as ways of viewing, presenting, valuing, or ordering this non-textual substratum.

This has generally been a fruitful way of proceeding, but as my description may already have suggested, it involves an operation which can certainly be questioned: the heuristic definition of a 'true sequence of actions' which narrative discourse is then said to present. The analyses of narrative which I propose to discuss bracket the question of point of view and implied narrator and treat the fabula itself not as a given but as a tropological construct. If one wished to identify this sort of analysis by its theoretical allegiances, one might cite Nietzsche's tropological deconstruction of causality ( De Man, 1974, 1975) or Kenneth Burke's account of narrative as the tropological 'temporalizing of essence' ( 1969: 430-40) [ . . . ]

Paul de Man's view of narrative as the expansion or literalization of tropological structures ( 1977) and J. Hillis Miller's account of the way narratives claim the status of history for their plots and then show history to originate in an act of discursive interpretation ( 1974) also belong to this general approach to narrative, but I am interested in more restricted analyses which explicitly identify fabula not as the reality reported by discourse but as its product.

To illustrate the kind of problems and issues involved, let us start with a familiar example, the story of Oedipus. Traditional narrative analysis would identify the series of events which constitute the action of the story ( Oedipus is abandoned on Mt. Cithaeron, rescued by a shepherd, raised in Corinth; he kills Laius at the crossroads, answers the Sphinx's riddle, marries Jocasta, seeks the murderer of Laius, discovers his guilt and blinds himself), and would describe the order and perspective in which these events of the plot are presented in the discourse of the play. It assumes, in other words, that these events constitute the reality of the story and it then asks about the significance of the way in which they are presented. In the case of Oedipus as in many other narratives, of which the detective story is the most common instance, the discourse focuses on the bringing to light of a crucial event, which the story identifies as a reality which determines significance. Someone killed Laius, and the problem is to discover what in fact happened at that fateful moment in the past.

This way of thinking about the play is clearly essential to its power, but one can also argue that this supposed event, Oedipus's slaying of Laius, is not something given as reality but is produced by a tropological operation, the result of narrative requirements. Once we are well into the play, it is clear to us that Oedipus must be guilty, otherwise the tale will not work at all. And this is not simply a matter of the reader responding to a peculiar aesthetic logic. Oedipus too feels the force of this logic. It has been prophesied, after all, that he would kill his father. It was prophesied that Laius would be killed by his son. Oedipus admits to having killed an old man at about the right time and place. So when the shepherd reveals that Oedipus is in fact the son of Laius, Oedipus leaps to the conclusion, and every reader leaps with him, that he did in fact kill Laius. He leaps to this conclusion not on the basis of empirical evidence or eyewitness accounts (the only witness has told a story that is incompatible with Oedipus's guilt: that there were many murderers, whereas Oedipus claims he acted alone), but is compelled rather by meaning, by the interweaving of prophesies and the demands of narrative coherence. Instead of saying, therefore, that there are events which took place and which the play reveals in a certain order and with certain detours, we can say that the crucial event itself is a product of the demands of signification. Instead of event determining meaning and meaning being the result or effect of a crucial event, it turns out that meaning is the cause of the event, the cause of its cause, in a tropological operation that can be assimilated to metonymy ('substitution of cause for effect or effect for cause').

Oedipus becomes the murderer of his father not by a violent act that is brought to light but by deeming this act to have taken place, by bowing to the demands of narrative coherence. Readers cannot escape this process either: the text persuades us that this event must have taken place. Although in theory the deed ought to be the cause of Oedipus's guilt, the play makes possible an alternative reading in which cause and effect are reversed and guilt is what produces the deed. And it is essential to the tragic force of the play that Oedipus should take this leap, embrace guilt and deem the act to have taken place rather than, say, resist the logic of signification and deny his guilt until actual proof is obtained. In identifying this alternative logic in which the event is not a cause but an effect of theme, one is helping to account for the force of the narrative.

But it is certainly not the case that in describing the play in this way we have replaced a deluded or incorrect model of narrative by a correct one. On the contrary, it is obvious that much of the power of Oedipus depends on the assumption that there is a prior fact which occupies a determinate place in the causal sequence of actions but which the narrative perspective only gradually reveals. It is essential to believe that Oedipus's innocence or guilt is determined by an event in the past which has already taken place but which has not yet been revealed. On the other hand, as we have already suggested, the tragic force of the ending depends on precisely the contrary logic whereby Oedipus posits the act on the basis of a structure of signification and whereby he and the readers are convinced that he did kill his father not by decisive testimony of witnesses (the witness is never even asked the crucial question) but by the demands of signification.

The play requires, then, a double analysis, a reading in two registers, one of which assumes the priority of acts to their narrative representation or presentation, the other which sees plot as a tropological product of narrative requirements. If Oedipus seems an unusual case in that it contains a possible uncertainty about a central event in the plot, let us consider an example from a very different period and genre, George Eliot Daniel Deronda, as analysed in a recent article by Cynthia Chase ( 1978). Deronda, the adopted son of an English nobleman, is a talented, sensitive young man, moving in good society, who has been unable to decide on a profession. He happens to rescue a poor Jewish girl who was trying to drown herself, and in searching for her family later, he meets her brother Mordecai, an ailing scholar with whom he begins to study Hebrew. He develops an intense interest in Jewish culture, falls in love with Mirah, the girl he has saved, and is accepted by Mordecai and others as a kindred spirit.

At this point Deronda receives a summons from his mother who, obeying her father's injunction, reveals to him the secret of his birth: he is a Jew. The novel emphasizes the causal force of this past event: because he was born a Jew, he is a Jew. Origin, cause, and identity are linked in a implicit argument that is common to narrative. With

the revelation of Deronda's parentage it is implied that his present character and involvement with things Jewish have been caused by his Jewish origin.

But on the other hand, as Chase notes, the novel's account of Deronda's situation has made it clear to the reader that the progression of his destiny and thus of the story positively requires that he turn out to be Jewish. Suspense is focused on his relationship with Mirah and Mordecai, so that something like this revelation is required for the resolution of the plot; and Mordecai has explicitly stressed his faith that Deronda is Jewish. Thus it seems that Deronda's Jewish parentage is something that can be deduced first from his identity -- his qualities and behavior as presented in his relations with Mordecai and Mirah -- and second from the patent strategy and direction of the narrative. 'The revelation of Deronda's origins therefore appears as the effect of narrative requirements. The supposed cause of his character and vocation (according to the chapters recounting the disclosure), Deronda's origin presents itself (in the light of the rest of the text) rather as the effect of the account of his vocation: his origin is the effect of its effects' ( Chase, 1978: 218).

It is important to stress here that, as in the case of Oedipus, there is no question of finding a compromise formulation which would do justice to both presentations of the event by avoiding extremes, for the power of the narrative depends precisely on the alternative use of extremes, the rigorous deployment of two logics, each of which works by excluding the other. It won't do to say, for example, that Deronda's involvement with Judaism is partly but not completely the result of his birth, and that the revelation of his birth is therefore in part an explanation and in part a narrative fulfillment. This sort of formulation is wrong because the power of Eliot's novel depends precisely on the fact that Deronda's commitment to Judaism and idealism, instead of to the frivolous society in which he has been raised, is presented as a free choice. To have exemplary moral value it must be presented as a choice, not as the ineluctable result of the hidden fact of parentage. It also must be presented as wholehearted, not as a dilettantish dabbling which would then be transformed into commitment by revelation of the fact of birth. The novel requires that Deronda's commitment to Judaism be independent of the revelation of his Jewishness -- this is thematically and ethically essential -- yet its account of Jewishness does not allow for the possibility of conversion and insists on the irreplaceability of origins: to be a Jew is to have been born a Jew. These two logics, one of which insists upon the causal efficacy of origins and the other of which denies their causal efficacy, are in contradiction but they are essential to the way in which the narrative functions. Once logic assumes the primacy of events; the other treats the events as the products of meanings.

One could argue that every narrative operates according to this double logic, presenting its plot as a sequence of events which is prior to and independent of the given perspective on these events, and, at the same time, suggesting by its implicit claims to significance that these events are justified by their appropriateness to a thematic structure. As critics we adopt the first perspective when we debate the significance of a character's actions (taking those actions as given). We adopt the second perspective when we discuss the appropriateness or inappropriateness of an ending (when we debate whether these actions are appropriate expressions of the thematic structure which ought to determine them). Theorists of narrative have always, of course, recognized these two perspectives, but they have perhaps been too ready to assume that they can be held together, synthesized in some way without contradiction. It is precisely this contradiction which will often manifest itself as a moment in the story that seems either superflous or else too neat, which recent work on narrative has brought to the fore, stressing its importance to the rhetorical force of the text.

Though my examples so far have been Oedipus Rex and Daniel Deronda, this double logic is by no means confined to fictional narratives. In several interesting articles Peter Brooks has discussed narrative in Freud, describing a complex situation ( 1977, 1979). On the one hand, Freudian theory makes narrative the preferred mode of explanation. Psychoanalysis does not propose scientific laws of the form 'if X, then Y.' Psychoanalytic understanding involves tracing a phenomenon to its origin, seeing how one thing leads to another. Freud's case histories are themselves often narratives of his conduct of the case, but, more important, he explains a neurosis or psychosis by reconstructing a plot, by telling what happened. Like Oedipus and Daniel Deronda, Freud's narratives lead to the revelation of a decisive event which, when placed in the true sequence of events, can be seen as a the cause of the present situation.

In Totem and Taboo Freud tells of a decisive historical event in primitive times: a jealous and tyrannical father, who kept all the women for himself and drove away the sons as they reached maturity, was killed and devoured by the sons who had banded together. This 'memorable and criminal deed' was the beginning of social organization, religion, and moral restrictions, since the guilt led to the creation of taboos. This historical event, Freud claims, remains efficacious to this day. We inherit and repeat the wish if not the actual deed, and the guilt which arises from this wish keeps the consequences of the deed alive in an unbroken narrative.

But clearly if guilt can be created by desires as well as by acts, it is possible that the originary act never took place. Freud admits that the remorse may have been provoked by the sons' fantasy of killing the father (by the imagination of an event). This is a plausible hypothesis, he says, 'and no damage would thus be done to the causal chain stretching from the beginning to the present day' ( 1950: 16). Choosing between these alternatives is no easy matter; however, he adds, 'it must be confessed that the distinction which may seem fundamental to other people does not in our judgment affect the heart of the matter.' [ . . . ] Emphasis on event and emphasis on meaning give the same narrative. But once again, one cannot fail to wish to choose, and Freud does: primitive men were uninhibited; for them thought passed directly into action. With them it is rather the deed that is the substitute for thought. And this is why, without laying claim to any finality of judgment, I think that in the case before us it may be assumed that 'in the beginning was the Deed'' ( 1950: 161).

A safe assumption, perhaps, but safe because it is so equivocal. Freud here starts with the fantasy and asserts that for primitive men the deed was a substitute for the fantasy. The deed truly took place, he claims, but his formulation prevents one from taking the deed as a given since it is itself but a substitute for the fantasy, a product of this primal fantasy. And in claiming that in the beginning was the Deed, Freud refers us not to an event but to a signifying structure, another text, Goethe Faust in which 'deed' is but a substitute for 'word'. Quoting Goethe in asserting an originary deed, Freud cannot but refer us to an originary word: the word that was in the beginning and the original authoritative word of Scripture. The two perspectives, in the beginning was the Deed and in the beginning was the Word, are certainly in conflict, and ethical concerns demand a choice, but as Freud's text shows, even when one tries to choose one does not escape the alternative one tried to reject.

I have been offering examples of an approach to narrative which, unlike studies of point of view that have dominated American narrative theory in the past, does not assume the primacy of event so as to focus on means of presentation but rather explores the complex interaction of two modes of determination which both seem necessary to narrative but which do not give rise to a harmonious synthesis. The examples I have used have been literary and theoretical narratives and I would like to conclude with a brief discussion of another species, what William Labov calls 'natural narrative.'

In his studies of the black English vernacular, Labov became interested in the narrative skills displayed by adolescents and preadolescents in responding to questions like 'Were you ever in a fight with somebody bigger than you?' followed by 'What happened?' if they answered 'Yes.' Labov's formal analysis of the elements of these stories ( 1967, 1972) begins by assuming the primacy of events, defining a narrative as a way of recapitulating past experience by matching a series of clauses to a sequence of events ( 1972: 360-1). However, what Labov discovers as a result of this orientation is that 'there is one important aspect of narrative which has not been discussed -- perhaps the most important element in addition to the basic narrative clause. That is what we term the evaluation of the narrative: the means used by the narrator to indicate the point of the narrative, its raison d'tre' ( 1972: 366). Indeed, it becomes clear that for Labov the narrator's primary concern is not to report a sequence of events but to tell a story that will not be seen as pointless. 'Pointless stories are met (in English) with the withering rejoinder, 'So what?' Every good narrator is continually warding off this question; when his narrative is over it should be unthinkable for a bystander to say, 'So what?'' ( 1972: 3 66)).

Labov's narrators prove skilled at warding off this question, at constructing their narratives in such a way that the demands of significance are met and the story is perceived as narratable, as 'reportable.' The question of whether any given story is being told primarily in order to report a sequence of events or in order to tell a tellable story is of course difficult to decide, but the ethical and referential lure of stories makes listeners want to decide (is that the way it really happened or is he just trying to impress us?). Labov avoids this question, either because he thinks it makes no difference to his analysis or because he thinks that the two projects can and should coincide, evaluative clauses being added to narrative clauses to produce a good story. As long as narrative clauses can be distinguished from evaluative clauses it is possible to maintain the view that a narrative is a sequence of clauses reporting events (which may be true or false) with clauses added which evaluate the event (and which are neither true nor false), but when Labov comes to describe the evaluative devices, he notes that some of the most powerful evaluative elements are not comments external to the action but embedded in the action itself. One can emphasize the reportability of a story, for example, by narrating as an event an evaluative comment: 'And when we got down there her brother turned to me and whispered, 'I think she's dead, John!'' ( 1967: 39). Indeed, the evaluation 'may itself be a narrative clause' ( 1967: 37) as in 'I never prayed to God so fast and so hard in all my life!'

Labov's claim is certainly correct, that many clauses reporting action are in fact determined by the evaluative function, i.e., the attempt to make the story a truly tellable story and avoid the question 'So what?' But given this possibility, we find ourselves in an awkward situation. For any report of an action there is always the possibility that it should be thought of as evaluative, as determined by requirements of significance, rather than as a representation of a given event. Whichever option we choose, we have the same narrative, of course, and in this sense it may not matter, but if we are concerned with the force of the story, and tellers and listeners of natural narrative are especially concerned with the force of narrative, then we are invited to choose and cannot take for granted the harmonious reconciliation of these two functions or logics.

So even here, in these most unliterary of narratives, we find the same problematic relationship between the determinations of fabula and the discourse of sjuzhet. [ . . . ]

References

BOOTH W., 1961. Rhetoric of Fiction ( Chicago: Chicago University Press).

BROOKS PETER, 1977. ''Freud's Masterplot: Questions of Narrative,'' Yale French Studies 55/ 6: 280-300.

----- 1979 ''Fictions of the Wolfman,'' Diacritics 9. 1: 72-83.

BURKE KENNETH, 1969. A Grammar of Motives ( Berkeley: California University Press).

CHASE CYNTHIA, 1978. ''The Decomposition of the Elephants: Double-Reading Daniel Deronda.'' PMLA 93: 215-27.

----- 1979 ''Oedipal Textuality: Reading Freud's Reading of Oedipus,'' Diacritics 9. 1: 54-71.

CHATMAN S., 1978. Story and Discourse: Narrative Structure in Fiction and Film ( Ithaca, NY: Cornell University Press).

DE PAUL MAN, 1974. ''Nietzsche's Theory of Rhetoric,'' Symposium: 33-45.

----- 1975 ''Action and Identity in Nietzsche,'' Yale French Studies 52: 16-30.

----- 1977 ''The Purloined Ribbon,'' Glyph 1: 28-49.

FREUD SIGMUND, 1950. Totem and Taboo ( New York: Norton).

GENETTE GRARD, 1972. ''Discours du rcit,'' Figures III ( Paris: Seuil).

LABOV WILLIAM, 1967. ''Narrative Analysis: Oral Versions of Personal Experience,'' Essays on the Verbal and Visual Arts: Proc. of the Annual Spring Meeting of the American Ethnological Society ( Seattle: Washington University Press), pp. 12-44.

----- 1972 Language in the Inner City (University Park: University of Pennsylvania Press).

LUBBOCK P., 1957, Craft of Fiction ( New York: Viking).

MILLER J. HILLIS, 1974. ''Narrative and History,'' ELH 41: 455-73.