Fabula and fictionality in narrative theory

The distinction between fabula and sujet is, according to various commonsensical definitions, the distinction between what happens in a narrative and how it is told; narrative theory, however, has struggled to reconcile common sense with conceptual rigor. The terms themselves derive from Russian Formalism, but the basic opposition they articulate is much older--it is there in the Poetics (everyone agrees that sujet corresponds to Aristotle's muthos, but whether fabula is best equated with praxis, or logos, or holos rather depends on which theorist you are reading). Numerous alternative terms have been proposed since the Formalists, too, and later in this article I shall be referring to story and discourse (Seymour Chatman), and histoire, narration, and recit (Gerard Genette). Such terminological revisions have typically sought to give new inflections to the Formalist pair, or to make the cut in different ways, but a sense of conceptual continuity dominates nonetheless, and if anything this sense becomes stron ger as the terms proliferate. In current usage there is no clear distinction between fabula and, for example, story, despite the latter's Structuralist pedigree (indeed the term "story" itself is blurred both by its nontheoretical currency and by its association with E. M. Forster's rather different contrast between story and plot, where the distinguishing criterion is plot's relative emphasis upon causality). My use of fabula here is meant only to invoke the consensus underlying the terminology: a shared but variously articulated sense of something fundamental to the understanding of narrative. I don't want to exclude, in the first instance, any of the ways in which this elusive something has been conceived; my purpose, though, is to work towards a viable concept by a process of elimination.

The theoretical problems were already apparent in the Formalists' own usage. They had reacted against realist poetics by inverting the priority of content over form, so that in their vocabulary "device," rather than "material," was privileged as the essence of art; and in the same spirit they treated fabula as just a foil for the literary effects of sujet. Boris Tomashevsky goes so far as to suggest that fabula may exist outside the realm of the work altogether: "Real incidents, not fictionalized by an author, may make a story [fabula]. A plot [sujet] is wholly an artistic creation" (68). He says "fictionalized" rather than "narrativized" (a point to which I'll return), but he seems to assume that fabula may precede all narrative-- may be found ready-made in real life. The motive, clearly enough, is to exclude the mere material of narrative creativity from the domain of art, but the consequences are paradoxical because any attempt to locate fabula outside the narrative domain is bound to deny it the specifici ty and integrity that constitute a fabula as such. The totality that is a "world" lacks, of itself, any principles of organization (spatially or temporally) by which to delimit a specifically narrative structure for sujet to manipulate. Whatever view we may wish to take upon the actual relations existing between the multitude of real events, the isolation of any particular sequence is already the intervention of narrative artifice. This is not really a matter of the logical or evaluative priority of either term, but simply of the linear, developmental nature of narrative not provided for by Tomashevsky's off-the-peg notion of real world fabula--nor, for that matter, by fictional worlds approaches to fiction. Fabula must be in some sense storied: it can't be understood as simply the world of the story, whether actual or fictional, because that would strip it of any specific relevance to narrative.

For theorists who reject the Formalists' sense that fabula is innocent of artifice, though, a predictable consequence is that the distinction between fabula and sujet starts to collapse. This collapse can be seen happening in an early critique of formalist theory, The Formal Method in Literary Scholarship, that begins by accepting the terms, if not the basis, of Tomashevsky's distinction: "although we can separate story [fabula] from plot [sujet] as the formalists understand it, the story itself is, nevertheless, artistically organized" (Bakhtin/Medvedev 113). This move soon modulates into a rather abrupt declaration that the distinction itself is redundant: "It is impossible to distinctly separate the two and to no purpose to do so. An orientation toward plot, i.e., toward the definite and actual development of the work, is necessary to master the story. Even in life we see the story with the eye of the plot [...] Thus story and plot are essentially the same constructive element of the work" (139). The lurch from unsustainable dualism to reductive monism here is one that recurs in later arguments about the nature of narrative. Perhaps the clearest instance is an exchange between Seymour Chatman and Barbara Herrnstein Smith in the pages of Critical Inquiry (Chatman, "Reply"; Smith). In Smith's critique of narratology, Chatman stood as the principal American synthesizer of Structuralist approaches to narrative, and his terms, story and discourse, were the focus of her argument. Structuralist narratology had readily appropriated the Formalists' dualistic perspective, but their tendency to privilege sujet over fabula, device over material, was now reversed: story, though necessarily seen through the prism of discourse, had become essential and primary. From the Structuralist perspective, some such distinction as that between fabula and sujet was fundamental to narrative theory--was in fact definitional of narrative insofar as it could be understood as both a coherent discursive category and a structure capable of ar ticulation in different media. Fabula or story, in this context, provided the deep structural logic underlying all the contingencies of a narrative's manifestation in the form of any particular sujet or discourse. For Smith, our sense that there may be different versions of a narrative did not require that they should be referred to a common structural basis, as discursive variants of the same story. Against the deep structural notion of story, she argued that the similarity between narratives could not be conceived in structural terms without simply producing another version: not story, but more discourse. Chatman's response, in the face of this monism, was to reassert his dualistic perspective; but he did little to address Smith's objections on their own ground. He suspected her of confusing story and plot summary, and he condemned her exclusively linguistic frame of reference, but above all he insisted, in the face of all theoretical objections, upon the sheer utility of the concept of story. Yet if the ba lance of the argument was on Smith's side, Chatman has in some sense been vindicated by the subsequent history of the concept: story has indeed been found too useful to abandon, even by some critics who share Smith's sense of its theoretical illegitimacy.

In many respects, the Structuralist enterprise is now as much a part of the history of literary theory as Formalism: narratology's figuration of narrative as a language (in accordance with Structuralism's linguistic paradigm) has proved unsustainable, and its decline has much to do with criticism of the opposition of story and discourse. But the distinction persists, and in certain respects--I want to suggest--with good reason. My claim is that the concept of fabula, appropriately understood, is not only useful to narrative theory, but crucial to any account of the interpretation of fiction. In order to make good this claim, I need to distinguish my understanding of fabula from the many respects in which it proves unsatisfactory, including its associations with Structuralist narratology (under the guise of story and other terminological variants). I am not proposing to resurrect the Formalist perspective except in the one respect that arises from my specific concern with the relation between fabula and fictio nality. The Formalists were concerned fundamentally with the artifice of literature, precisely where they located both literature's specificity as an object of study and its value as a cultural phenomenon. So, when Victor Shklovsky invokes the concepts of fabula and sujet, it is in order to celebrate the self-conscious artifice of Sterne's games with the telling and the told in Tristram Shandy (25-57). When Tomashevsky uses the distinction, it is to articulate a model of literary thematics as an organizing principle of narrative, in opposition to representational logic: he conceives of the primary materials of fiction not as events, but "motifs" (61-95). The Formalists perceived that fictional narrative is primarily driven by the dictates of art (however they might be conceived), and that its conformity to a logic of representation is always in some sense dissimulated. The implications of this specific interest in fictionality have been obscured in the subsequent history of narrative theory, however, precisel y to the extent that the artifice involved has come to be understood as an attribute of narrative in general. The Formalists' focus upon "art" was not specific to narrative, and so inhibited their efforts to explain narrative representation; but the narratologists' more rigorous isolation of the category of narrative was all too often oblivious to the issue of fictionality, despite their own almost exclusive attention to fictional narratives. In fact the habit of narrative theory in general has been to use fictional narrative as its paradigm, but nonetheless to appropriate a nonfictional, and philosophically realist, notion of narrative events (an example from a solidly anti-Structuralist quarter would be the orientation of fictional-worlds approaches to fictional reference). As a necessary first step in redeeming the theoretical value of fabula, it must be detached from the concept of event.

Events, and reference to them, seem to present an obvious basis for fabula's relation to sujet. Definitions of the minimal criteria for narrative are often couched in terms of specified relations between events, understood as the basic units of the sequence that underlies whatever arrangements and selections may constitute the sujet of a given narrative. There is little consensus among theorists as to the detail, but most of the disagreements are relatively trivial consequences of a tendency to overspecify the nature of the required relations between events. One issue, though, is fundamental: must a narrative consist of at least two events, or may only one suffice? I don't wish to adjudicate, but only to note that the possibility of disagreement upon this point implies that the concept of event may be understood in two contrasting ways. One position treats the event externally, as a singularity, so that narrative temporality only emerges in the passage from one to the next; the other locates narrative within the event, in the transformation it marks between before and after. The mutual presupposition entailed by these two interpretations of event doesn't admit of resolution without recourse to metaphysical argument, but that doesn't matter here, because I only want to show that reference to events cannot in any case provide for a definition of narrative or a concept of fabula. In either case, the concept of event begs the question of narrative, for narrative either slips between or continues to lurk within event; only the coexistence of two different senses of event allows the evasion to go unnoticed.

Reference to an event is not narration of an event; equally, there is no theoretical dividend to be gained from resolving the narration of an event into sequential reference to (sub) events. The idea of the event cannot, in itself, provide for the underlying structure of narrative discourse: on the contrary, the event is always susceptible to narrative decomposition. Implicit in every event referred to in a narrative is its internal chronology; that is to say, its status as a process with a beginning and end--as itself latent narrative. In the absence of any conceptual limit to this regressus, only narration itself, defined nontautologically as the delineation and delimitation of process, can confer narrative status. Any narrative analysis that uses events as its fundamental units is necessarily regarding them as quanta, but not according to any criteria other than those supplied by the sujet: that is, typically, linguistic criteria. The view of the event as an integral unit is influenced by the integrity of the sentence, or of the verb phrase: as soon as the question is transferred to a nonverbal narrative medium such as film, it is impossible to conceive of the event other than in durational terms. Narration in any medium represents the event in process, not by invoking it as a narrative unit; and it is only as process that the event is constituted as narrative at all.

The only solid ground available, then, lies on the plane of sujet, not fabula: it is the categorical distinction between reference and narration. This itself, in the limit case, comes down to a choice of interpretative stance: for example, nothing else can determine whether many early single-shot films are descriptive or narrative in mode. A common genre of early actualities entertained audiences with scenes of moving water (waves, rough seas, etc.): their point, of course, was to exhibit the new medium's remarkable powers of descriptive reference. But the situation is more ambiguous in instances like the 1897 Lumiere travelogue of Niagara Falls, in which a group of people congregate at the rail on the far bank. If the interpretative focus is the Falls, the mode remains descriptive (this is undelimited process); but if the action on the far bank is the center of attention, the film becomes (minimally) narrative. The "event," here, has no status independent of the discourse and its interpretation. Because narr ative cannot be defined in terms of reference to events, neither can fabula be construed as the referent of sujet.

Beyond its association with narrative events, the most prevalent assumption about fabula is that it is unmarked by the distortions of chronological sequence that may characterize sujet. The Formalists were immensely interested in works (like Tristram Shandy) that exploit the relation between the order of telling and the order of occurrence, and this preference undoubtedly encouraged an equation between fabula and the zero degree of chronology. But since a narrative may just as well be told in straightforwardly chronological fashion, such a sense of fabula runs an immediate risk of losing its analytical force. In fact, as Meir Sternberg has shown in detail, chronological narration is far from neutral in its temporal organization. Sternberg himself therefore finds it useful to relate even the most straightforward sujet to a chronological notion of fabula, on the grounds that the dual-level model of fabula and sujet captures the distinguishing feature of narrative effects, their "interplay between times": fabula 's development in mimetic time and sujet's discourse in communicative time ("Telling in Time [II]" 519, 531). Yet at the same time his attentions to the artfulness of chronological narrative problematize the notion of mimetic time itself: it becomes increasingly clear that chronological sequence as such can't serve to define fabula, because it can't be reduced to a degree zero unmarked by temporal and perspectival selection. There is never an absolute chronology of events, not because they cannot be put in temporal sequence but because any such sequence always remains inflected by interpretative choices: most fundamentally, the determination of what shall count as the events, and what the transition from one to the next (Zeno thrived upon just this conceptual indeterminacy).

Sternberg's emphasis upon narrative effects is very properly part of his argument for a functionalist approach to narrative: in the absence of an objective norm of chronology, categorical differences in narrative treatment can only be described as differences of effect ("Telling in Time [II]" 497). But these effects often resist the dual temporal order of fabula and sujet itself. For example, he notes the pervasiveness of simultaneity in narrative: that is, the representation of concurrent events necessarily consecutive in the telling because of the linear nature of sujet. In such cases the sujet arrangement of events is an imposition of sequential order, not the reordering of any fabula-like chronological sequence ("Telling in Time [I]" 941). A related possibility is noted by Genette in Narrative Discourse, where he observes that a narrative may include "events not provided with any temporal reference whatsoever. [...] To be unplaceable they need only be attached not to some other event (which would require the narrative to define them as being earlier or later) but to the (atemporal) commentarial discourse that accompanies them" (83). In other words, the organizational principle of the narrative in such circumstances cannot be specified in relation to the chronology of events narrated, but only at the level of the narration. Genette's examples, of course, come from A la recherche du temps perdu, but this phenomenon is far more common (on a more modest scale) than narrative theories have recognized: it won't do to treat it, with Genette, as anomalous. Even one of the classic instances of the sujet's disruption of fabula chronology, the in medias res opening, cannot finally be accounted for in these terms. The effect of such openings clearly relates to all the questions begged by the immediate narrative action; but in the absence of any prior knowledge of the whole narrative (which distinguishes such circumstances as second readings, or stories drawing upon well-defined mythical or folkloric traditions), this dis tinction can have nothing to do with a sense of disjunction between sujet and fabula chronology, because it precedes the reader's awareness of any such chronology. While Sternberg's discussion of such matters effectively demonstrates the need to ground theoretical approaches to narrative ordering in effects that "variously manifest themselves on the narrative surface" ("Telling in Time [II]" 497) rather than in the sujet's relation to an underlying fabula chronology, I would want further to urge that the representational priority of fabula (as the virtual event-sequence "reconstructed" from sujet information) is itself undermined by such arguments.

The relation between narrative and temporality itself also resists treatment as an aspect of sujet's manipulations of fabula. Perhaps the least successful part of Genette's treatment of narrative time is the notion of narrative duration, or pace: he describes this variable by appealing to the idea of speed, because it concerns the relationship between a spatial dimension (of "recit" or sujet--he confines himself to written media) and a temporal dimension (of "histoire" or fabula), and so may be calibrated in lines (or pages) per minute (or day, or year) (87-88). But the account is unconvincing because narrative duration really just elaborates upon the old distinction between scene and summary, and doesn't pertain to temporality so much as to the quantity of narrative information. Genette's principles of fabula chronometry do little to account for the narrative sense of time, not least because at bottom temporality and narrative are interdependent concepts. Paul Ricoeur, in Time and Narrative, dedicates three volumes to the proposition that "time becomes human time to the extent that it is organized after the manner of a narrative; narrative, in turn, is meaningful to the extent that it portrays the features of temporal experience" (1: 3). To the extent that narrative is already inherent in our sense of temporality, it is fruitless to appeal to that sense as the fundamental ground upon which narrative is built. Ricoeur, commenting on the circularity of his own argument, observes that "the manifest circularity of every analysis of narrative, an analysis that does not stop interpreting in terms of each other the temporal form inherent in experience and the narrative structure, is not a lifeless tautology. We should see in it instead a "healthy circle" in which the arguments advanced about each side of the problem aid one another" (1: 76). This is, I think, an apt formulation; but it clearly denies precedence to either side, and so disqualifies any model of fabula that claims such precedence by invoking temporality as its benchmark.

For some the concept of fabula is better characterized by appeal to the underlying causal logic of narrative, rather than mere chronology. Certainly Tomashevsky emphasized the association between fabula and causality from the Formalist point of view, and others have followed. But such an approach to fabula is hardly viable, even leaving aside the Humean argument that, as Adams says, "causality is not in the world we experience but in the discourse we use to represent the world" (151). It is no more possible to define the causal linkage between a series of events than it is to definitively identify those units called events in themselves. Our causal understanding of a narrative certainly cannot be restricted to the causal relations explicitly stated: no narrative can ever be fully explicit about the logic linking the events it represents. Even the most intrusive novelist will necessarily leave a great deal implicit, and in nonlinguistic media the notion of explicitly causal relations is almost unintelligible. The interpretative activity this requires of the reader is not of a kind that strips away the discursive surface to expose an innate logic in the particulars of the narrative events: rather it is a process of narrative supplementation. A narrative's silences co-opt the reader's interpretative collaboration, and so serve a rhetoric of verisimilitude rather than merely resting upon a subtext of causal necessity. Causal explanation, whether given in the sujet or inferred from it, does not invoke an underlying, discourse-independent logic, but produces it: it is only possible to build a sense of a narrative's representational causality in the terms dictated by the logic of the sujet itself. Indeed it might be said, to adopt a formula from Roland Barthes, that the most elementary function of narrative discourse is to produce and make literal the fallacy of post hoc ergo propter hoc (94).

Perhaps the most fundamental weakness in the concept of fabula though, is betrayed by the Formalists' bias towards questions of temporal ordering at the expense of perspectival issues. Point of view has always been central to narrative theory, but relatively little attention has been paid to the problems it raises about the ground against which perspectival restrictions are perceived. Fabula, in order to function as the untransformed substructure beneath the adopted point of view of the sujet, would need to be innocent of all perspective. If fabula is, as it must be, narrative in form--not world, but story--then this requirement is flatly impossible: the difference between an undifferentiated mass of congruent facts or events and a sequence exhibiting the coherence of a story can only be the product of perspectival criteria. Which is to say, again, that fabula cannot be logically prior to sujet. The use of the term "omniscience" in novel criticism should not be allowed to confuse the argument here: it can't p rovide a model for undistorted narrative, because it could not, taken literally, offer any narrative at all (it would be overwhelmed by multiplicity and simultaneity). "Omniscient narration," which is in any case a discursive category and belongs to sujet, is only omniscient in respect of its potential liberties with point of view, not in its actualization. It might be further argued that chronology and causal sequence are themselves best thought of as ultimately perspectival issues, inasmuch as they are aspects of the question of narrative unity. A sequence of events is constituted as a sequence by decisions about relevant and irrelevant consequences, and the degree to which relations of consequence should be pursued and elaborated. That there are no sequences of events independent of such choices alone makes it illegitimate to privilege chronological or causal sequence as an unmarked, unmediated substratum of narrative. The issue of perspective is not merely presentational and cannot be confined to sujet: i t goes all the way down. This amounts to a kind of Heisenberg uncertainty principle for narratologists: there is no way to conceive of narrative events that does not in itself affect them.

Up to this point I have barely acknowledged the issues raised by the several tripartite variations upon the distinction between fabula and sujet. Of these, Genette's terms "histoire," "recit," and "narration" are representative. He sets out his wares in this way: "I propose [...] to use the word story [histoire] for the signified or narrative content [...,] to use the word narrative [recit] for the signifier, statement, discourse or narrative text itself, and to use the word narrating [narration] for the producing narrative action and, by extension, the whole of the real or fictional situation in which that action takes place" (27). This scheme raises the issue of the narrating instance--which is to say the mediation of a narrative by a represented narrator (because for Genetle the narrating instance in fiction is always represented, even where this representation consists of nothing but the bare existence of an unpersonified narrator). By promoting the narrating instance in this way, Genette is merely formal izing some widely held assumptions. But in doing so he confirms the suspicion, as Fludernik argues, that "the story vs. discourse opposition seems to repose on a realist understanding of narration" (334). This has implications for both sujet and fabula: with respect to sujet, Genette' s representational model of narration is an issue on which, elsewhere, I have commented at length ("Who Is the Narrator?"); with respect to fabula, the model implies that fabula is always literal in relation to its own sujet. In other words, the distinction between the (complete, objective) fabula and the (partial) sujet of a fictional narrative is negotiated through a narrator (agent of the narrating instance) who, however incompletely or even unreliably, is communicating fictional facts. The inadequacy of this model is apparent if we consider, for example, a dream narrative. In such a context it is distinctly odd to state, as one critic does with reference to The Wizard of Oz, that "the hours Dorothy and her entourage spend wa lking on the Road are omitted" (Bordwell 109). Which hours are these? In the literal world of the film there are no such hours, because this is part of a dream, and no more accountable to realist temporal logic than to realist attitudes towards witches, scarecrows, lions and the rest. To take Oz literally would be to disregard our recognition of the correspondences between Oz characters and Kansas characters, and the extent to which the Oz narrative is driven by events in Kansas, above and beyond any representational logic of its own. Dream narrative does not submit to the literalism of Genette' s version of fabula because its structural logic is not founded upon dream facts, but dream meanings. And although it is superfluous to my argument at this point to say so, in this respect dream narrative may be thought of as merely a special case of the situation for all fictional narrative.

If the intermediate concept of "narration" does little to help the conventional distinction between fabula and sujet, further problems arise in connection with the actualizing medium itself. One of the most frequently reiterated arguments for a concept of fabula is that it is needed to explain the possibility of transposing the same narrative into another medium. A film and a novel, despite the inevitable changes in manifestation (or sujet) imposed by the capabilities of their different media, may both tell what is recognizably the same "story"; and, in Chatman's words, "This transposability of the story is the strongest reason for arguing that narratives are indeed structures independent of any medium" (Story and Discourse 20). In saying so he is only echoing a fundamental tenet of such Structuralist forbears as Claude Bremond and Roland Barthes. But is deep structure really the only way of accounting for this? To say that a film and novel have the same fabula, or story, is to say that the same plot summary would serve for both, although of course a plot summary itself cannot be regarded as "story" because it is a manifest narrative, with its own discourse features; and furthermore it is not unique--it is always possible to summarize a narrative in a number of ways. Nevertheless, to refer this equivalence back to a common deep structure adds nothing: any summary that worked for both novel and film would specify their sameness to just that extent, quite independently of any appeal to "story"; nor, in fact, can story itself be appealed to except in terms of some such exercise in summary. Such narrative sameness is interpretative, posterior (however much anticipated and provided for by authorial or directorial intention), and is negotiated in terms of actual versions, without recourse to an abstract structural congruence. The medium of plot summary itself, of course, is typically linguistic--but not inherently so: a film storyboard can perform the same function. Narratives are not so much structures independent of any medium, as structures common to several media: the medium, if nonspecific, is not extrinsic but necessary to the concept of narrative.

While the Formalists were sometimes happy to consider fabula as a manifest version of the narrative, most subsequent theorists (and certainly those within the Structuralist camp) have insisted that fabula is necessarily abstract. The role of abstraction in definitions of fabula is typically dictated by two considerations: that it is logically prior to any sujet, and that any manifest narrative necessarily already involves sujet. In Chatman's technical sense of the terms, "Story [...] exists only at an abstract level; any manifestation already entails the selection and arrangement performed by the discourse as actualized by a given medium" (Story and Discourse 37). It's hard to get at this sense of an abstraction of story or fabula that all too conveniently evades awkward questions about its specificity; and Chatman himself muddies the water by describing story content as "representations of objects and actions" that, as representation, would seem to be already discourse (24). A helpful attempt to illustrate t he idea has been offered by Edward Branigan, discussing story in relation to film narrative. He offers the analogy of an object represented in a film: "We know the object when we know how it may be seen regardless of the position from which it was actually seen. The object thus acquires an 'ideal' or 'abstract' quality. It should be mentioned that knowing how the object may be seen is very nearly imagining an object that is not in view at all" (15). This effectively conveys the idea that our knowledge of an object independent of any view is very like our knowledge of a fabula independent of any sujet. Except that, unlike an object, an event or sequence of events has no integrity (as event or as sequence, depending on whether you are taking an inside or outside view of the event) apart from that produced by its narration as such. Representation and narration are not interchangeable concepts, despite the dominance of a representational paradigm throughout the history of narrative theory and criticism. The notio n of fabula as the abstract plane of narrative, then, does little to redeem it: the sense of abstraction required is either inscrutable, or else accessible only in the form of a narrative version in spite of itself.

One other consideration seems to be involved in arguments for the logical priority of fabula: it is often associated with conceptual "raw material" in the genesis of the work. This was certainly true of the Formalists, for whom the articulation of a given fabula as sujet exclusively constituted the art of narrative creativity. Steinberg's account of fabula has sometimes taken this line too: "it may thus be viewed as the second-degree 'raw material' (postselected and straightforwardly combined narrative) that the artist compositionally 'deforms' and thus recontextualizes in constructing his work (mainly by way of temporal displacements, manifold linkage, and perspectival manipulations)" (Expositional Modes 8). Tellingly, such a view leads Sternberg on to countenance the possibility of a specific manifestation of fabula prior to the fully realized narrative, of which he takes James's notebook scenario for The Ambassadors to be an example: "This scenario being in fact a foreshortened fabula, the temporal and oth er presentational discrepancies between it and the finished product go to show that the concept of preexistent fabula, though indeed primarily a theoretical or reconstitutive rather than genetic model, may actually serve as 'raw material' and starting point for the process of artistic manipulation" (313n.). Yet the scenario is itself a narrative, and so ought itself to comprise both fabula and sujet: such preliminary narratives (whether extant or in the author's head) are not "raw," but already parboiled. Sternberg's concept of fabula here both encroaches upon that of sujet (since to sketch a narrative in outline remains a way of telling it) and requires a further fabula-like frame of reference to be projected beyond itself, producing an infinite regress--one, indeed, that might be equated with the intentionalist regress inherent in such an approach to creativity.

In general, a fundamental slippage between incompatible models of fabula occurs whenever accounts of its relation to the creative and interpretative processes are juxtaposed. Chatman, for example, discussing the necessary selectiveness of discourse in relation to story, invokes a notion of the "complete" account of a given course of events as the foil to that selectiveness: "each character obviously must first be born. But the discourse need not mention his birth" (Story and Discourse 28). He goes on to relate this never-realized complete account to his concept of story: "Thus story in one sense is the continuum of events presupposing the total set of all conceivable details, that is, those that can be projected by the normal laws of the physical universe. In practice, of course, it is only that continuum and that set actually inferred by a reader, and there is room for difference in interpretation" (28). By presenting discourse as narrative selection, Chatman unavoidably commits himself to a concept of the t otality from which that selection is made. But such a definition of story is unworkable (each character must presumably have parents), so he immediately qualifies it by appealing to interpretative practice, an appeal that makes story dependent upon readers' actual inferences from the discourse. This move distributes the concept of story between two senses, one genetic and ideal and the other interpretative and pragmatic, without acknowledging that the logical relation between story and discourse has been inverted in the process. Chatman speaks of the reader's inferences as "logically necessary," but this necessity derives from the general framework of understanding we bring to narrative interpretation, not from the specifics of a preexistent fabula (29). As the emphasis shifts from narrative articulation to narrative interpretation, so discourse, rather than story, takes logical priority as the delimiting framework of the narrative.

Chatman, then, shuffles between authorial and interpretative models of fabula in an attempt to reconcile two incompatible premises: that sujet is based upon fabula, and that fabula is derived from sujet. The author-oriented model insists upon the abstraction and logical priority of fabula as the basis for the selections and arrangements of the realized sujet; the reader-oriented model provides for the lack of any access to fabula except through sujet, by conceiving of fabula in pragmatic terms as an interpretative reconstruction. I want to suggest that this pragmatic model alone, once it has been detached from fabula's claim to abstract priority over sujet (that is, once it is recast as an interpretative construct rather than a reconstruction), can provide all the analytical power we need from a concept of fabula without falling prey to any of the conceptual problems I have been discussing. Fabula, in this view, is a function of interpretation: it doesn't strip the presentational features of sujet back to an objective narrative core, but reveals them as they are by imagining them otherwise. What that otherwise may be is largely contingent upon the orientation of a particular reading, but in general fabula simplifies sujet: it reduces it to the simplest terms consistent with the needs of the interpretation--most obviously to a chronological order wherever practical and useful. This reducing process is not for the sake of the residue, but the distillate: its concern is with the qualities of the narrative as given--the rhetoric of sujet. And though this view of fabula is centered upon reading, I have already suggested elsewhere that an interpretative model can serve equally well as an account of the creative process.

In what sense, then, does the concept of fabula continue to be valuable? Much of what I've said relates to the pervasiveness of narrative understanding in general, because that irreducible narrativity means that narrative knowledge is always characterized by the artifice of sujet, that artifice which is precisely what undermines the conventional idea of fabula. In fact, the persistence of a traditional concept of fabula is largely responsible for the common misconception that narrative artifice is inherently a "distortion"--in other words, that the narrative information can be conceived of in undistorted form. But the kind of artifice with which the Formalists were primarily concerned, and that interests me here, is of a different order from that which narrative theory has insisted is intrinsic to all narrative, fictional or nonfictional. Quite apart from any concept of fabula, historical narrative does have obligations to prior sources, whether narrative themselves (e.g., first-hand testimony) or nonnarrativ e (e.g., statistical records), and some theorists have sought to incorporate this into the fabula/sujet model. Dorrit Cohn, for example, suggests historical narrative may be distinguished from fiction by its adherence to a three level model of the form "reference/story/discourse" (779). Although this does nothing to resolve the contradictions in the concept of story, it does help to tease apart two different kinds of concern involved in narrative interpretation, and obscured by the tendency to use fiction as the model for general theories of narrative. Dominant in the nonfictional case are criteria for assessing a narrative's relation to the prior sources and intersecting narratives invoked by Cohn's term "reference." In this context the role of fabula is to serve the reader's evaluation of sujet in relation to (actual or hypothetical) alternative versions, with different sujet emphases, slants or omissions. At issue here is not the adequacy of fabula to source material (which would beg all the important ques tions), but the adequacy of sujet: fabula serves not as the uninflected story beneath the rhetorical manipulations of sujet, but simply as the means of throwing sujet rhetoric into relief. In nonfictional narrative it facilitates the dominant interpretative issue of comparison between actual or possible versions, in the context of adequacy to sources; but in the absence of (constrictive) sources or alternative versions it takes on a new significance. Fabula in the context of fiction is simply diagnostic of the rhetorical weight and distribution of salient values in the narrative itself. In fiction, fabula is both how we understand sujet per se, and how we understand its contingency (potentially, its unreliability), not in relation to facts or sources, and usually not in comparison with other versions, but with respect to its own disposition of values. The importance of the concept of fabula, then, is its direct link to rhetorical perceptibility, above and beyond issues of representational coherence: it is an interpretative exercise in establishing representational coherence only as a means to the end of this perceptibility. The role of fabula is enhanced in the interpretation of fiction, precisely because the role of rhetoric is enhanced: fabula serves to police the rhetoric of nonfictional narrative, but it serves much more fundamentally to explore the rhetoric of fiction, inasmuch as such rhetoric is the raison d'etre of fictionality.

A rhetorical (or functional) approach to narrative, and especially to fiction, anchors the concept of narrative in sujet rather than fabula (Sternberg, "Telling in Time [II]" 509-10, 513). Fabula is not so much an event chain underlying the sujet, as a by-product of the interpretative process by which we throw into relief and assimilate the sujet's rhetorical control of narrative information. Its integrity as a concept is not to be found in its relation to any given narrative, but to any given act of narrative interpretation. To recall the commonsensical definition of fabula and sujet with which I began, the value of the distinction may still be conceived in terms of a "what" and a "how," but with the terms reversed: sujet is what we come to understand as a given (fictional) narrative, and fabula is how we come to understand it. Our understanding, in other words, is not of "what happened"; it is of the weight and import of the narrative as actually told. The model of fabula I wish to affirm does not reject ab straction in favor of any specific manifestation (which would be no fabula at all): it merely inverts the logical priority of fabula to sujet. The reader's engagement with sujet does not enable the reconstruction of fabula, but its construction. Fabula is not independent of any sujet--it is entirely dependent upon sujet, is nothing other than the permutation and assimilation of sujet features into an ongoing interpretative version; neither is it ever realized in itself, since it functions only as an instrument of interpretation and is pursued only so far as the needs of the occasion demand. Fabula is always relative to and contingent upon both a given sujet and a specific act of interpretation. Such an understanding of fabula meets the need for which the concept has continued to be invoked despite some intractable theoretical problems, but avoids those problems by using a rhetorical rather than representational frame of reference. Reconceived in this way, fabula no longer trades off utility against conceptual incoherence: it is indeed a version, but not a prior version, and not extant; it is the reader's working version, a function of the process of interpretation, and a means rather than an end in itself. It serves not as the decoding of sujet, but as the triangulation of sujet features in order to gauge and respond to the rhetoric they embody. And in relation to fiction, its continued value as a concept lies in the fact that the possibilities of this rhetoric define the entire scope of fictionality.

Works Cited

Adams, Jon K. "Causality and Narrative." Journal of Literary Semantics 18 (1989): 149-62.

Bakhtin, M. M./P. N. Medvedev. The Formal Method in Literary Scholarship. 1928. Trans. Albert J Wehrle. Baltimore: Johns Hopkins UP, 1978.

Bal, Mieke. Narratology: Introduction to the Theory of Narrative. Trans. Christine van Boheemen. Toronto: U of Toronto P, 1985.

Barthes, Roland. "Introduction to the Structural Analysis of Narratives." 1966. Image-Music-Text. London: Fontana, 1977. 79-124.

Bordwell, David, and Kristin Thompson. Film Art: An Introduction. 1979. 5th Ed. New York: McGraw-Hill, 1997.

Branigan, Edward. Narrative Comprehension and Film. London: Routledge, 1992.

Bremond, Claude. "Le message narratif." Communications 4 (1964): 4-32.

Chatman, Seymour. Story and Discourse: Narrative Structure in Fiction and Film. Ithaca: Cornell UP, 1978.

____. "Reply to Barbara Herrnstein Smith." Critical Inquiry 7.4 (1981): 802-09.

Cohn, Dorrit. "Signposts of Fictionality: A Narratological Perspective." Poetics Today 11.4 (1990): 775-804.

Fludernik, Monika. Towards a "Natural" Narratology. London: Routledge, 1996.

Genette, Gerard. Narrative Discourse. 1972. Trans. Jane E. Lewin. Oxford: Basil Blackwell, 1980.

Lamarque, Peter, and Stein Haugom Olsen. Truth, Fiction and Literature: A Philosophical Perspective. Oxford: Clarendon P, 1994.

O'Neill, Patrick. Fictions of Discourse: Reading Narrative Theory. Toronto: U of Toronto P, 1994.

Prince, Gerald. Narratology: The Form and Functioning of Narrative. The Hague: Mouton, 1982.

_____. A Dictionary of Narratology. Lincoln: U of Nebraska P, 1987.

Ricoeur, Paul. Time and Narrative. 1983-85. 3 vols. Trans. Kathleen McLaughlin and David Pellauer. Chicago: U of Chicago P, 1984-1988.

Rimmon-Kenan, Shlomith. Narrative Fiction: Contemporary Poetics. London: Methuen, 1983.

Shklovsky, Victor. "Sterne's Tristram Shandy: Stylistic Commentary." Russian Formalist Criticism: Four Essays. Ed. Lee Lemon and Marion Reis. Lincoln: U of Nebraska P, 1965. 25-57.

Smith, Barbara Herrnstein. "Narrative Versions, Narrative Theories." Critical Inquiry 7.1 (1980): 213-36.

Meir Sternberg. Expositional Modes and Temporal Ordering in Fiction. Baltimore: Johns Hopkins UP, 1978.

_____. "Telling in Time (I): Chronology and Narrative Theory." Poetics Today 11.4 (1990): 901-48.

_____. "Telling in Time (II): Chronology, Teleology, Narrativity." Poetics Today 13.3 (1992): 463-541.

Tomashevsky, Boris. "Thematics." Russian Formalist Criticism: Four Essays. Ed. Lee Lemon and Marion Reis. Lincoln: U of Nebraska P, 1965. 61-95.

Toolan, Michael J. Narrative: A Critical Linguistic Introduction. London: Routledge, 1988.

Walsh, Richard. "Who Is the Narrator?" Poetics Today 18.4 (1997): 495-513.

_____. "The Novelist as Medium." Neophilologus 84.3 (2000): 329-45.

Richard Walsh (rmw8@york.ac.uk) is a lecturer in English and related literatures at the University of York. His publications include Novel Arguments: Reading Innovative American Fiction (Cambridge University Press, 1995); "Who Is the Narrator?" in Poetics Today (1997), "Why We Wept for Little Nell: Character and Emotional Involvement" in Narrative (1997), and "The Novelist as Medium" in Neophilologus (2000).

COPYRIGHT 2001 Northern Illinois University
COPYRIGHT 2003 Gale Group