In Reading for the Plot, [Peter] Brooks defines narrative as "the play of desire in time" (xiii) and identifies two sites for this play: first, the text itself, wherein desire to order compels the plot's unfolding; and, second, the space between text and reader, wherein the reader's desire for plot impels the reading (37-61). Analysis of these narrative desires involves seeing "the text itself as a system of internal energies and tensions, compulsions, resistances, and desires" (xiv). Like Paul Ricoeur in "Narrative Time," Brooks insists upon the temporal dimen­sion of narrative, on narrative's essential relation to time. I want to extend Brooks's "dynamics of narrative" by reintroducing the issue of space into a discussion of narrative, by considering narrative, in other words, as the play of desire in space as well as time.This essay presents the theoretical section of a much longer essay that includes a reading of Virginia Woolf's The Voyage Out, forthcoming in Kathy Mezei's Feminist Narratology. lt was originally presented at the lntemational Conference on Narrative in Nice, France, June 1991.

I define narrative most simply as the representation of movement within the coordinates of space and time.For my purposes here, I am not suggesting a masculine/feminine binary for time as space, as do Kristeva in "Women's Time" and de Lauretis in Alice Doesn't (143). See also Winnett's critique of Brooks's model. Here, I adapt M. M. Bakhtin's concept of the chronotope, by which he means the special form in which the "intrinsic interconnectedness of temporal and spatial relationships" is expressed in litera­ture (Dialogic Imagination 84). Invoking Einstein's theory of relativity, Bakhtin argues for the "inseparability of space and time" (84) and resorts repeatedly to spatial tropes in his analysis of various chronotopes. I also want to develop Julia Kristeva's adaptations of Bakhtin's spatial tropes in two early essays, "Word, Dialogue, and Novel" (1966) and "The Bounded Text" (1966-1967), both of which are included in her collection Desire in Language. Here, in introducing her concept of intertextuality, she advocates a reading practice based in the "spatialization" of the word along vertical and horizontal axes in an intertextual grid. In this essay, I will adapt her spatial tropes to suggest that we can read narrative by interpreting the text's horizontal and vertical narrative movements and intersections. Such interactions are events, I will argue, that take place at every moment in the text in a kind of interdependent interplay of surface and depth. Such moments may appear as juxtapositions, oppositions, conflations, convergences, or mirrorings of narrative coordinates. These moments in tum join to form a fluid "story" of a dynamic text ever in process, ever "narrated" by the reader.

KRISTEVA'S SPATIALIZATION OF THE WORD

For Kristeva, spatialization—with its attendant graphic tropes of coordi­nates, axes, trajectory, horizontal, vertical, surface, intersection, linearity, loop, dimension, and so forth—allows for the visualization of the text-in-process, the text as a dynamic "productivity," an "operation" (Desire 36-37). Spatialization does not mean the erasure of time by space, as it does for Joseph Frank, who, in his influential essay, "Spatial Form in Modem Literature," argues that avant­ garde narrative techniques in modem literature created an illusory effect of si­multaneity and unity. Rather, for Kristeva, spatialization constitutes the text as a verbal surface or place in which both space and time, synchrony and diachrony, function as coordinates for textual activity. Kristeva's earliest essays pose a critique of the static analysis of structuralism and a call for the identifìcation of textual process. Invoking Bakhtin, Kristeva identifìes this process as funda­mentally dialogic and intertextual—at the level of word, sentence, and story. Bakhtin, she explains, "considers writing as a reading of the anterior literary corpus and the text as an absorption of a reply to another text" (Desire 69). "Each word (text)," she continues, "is an intersection of word (texts) where at least one other word (text) can be read . . . . [A]ny text is constructed as a mosaic of quotations; any text is the absorption and transformation of another" (Desire 66). She graphs these intersections by identifying the text's "three di­mensions or coordinates" as the writing subject, the addressee, and exterior texts. What she calls the horizontal axis is a line drawn from writing subject across to the addressee, who is either a character to whom the speech is directed or, more generally, the reader. This horizontal axis represents the text as a trans­action between writer and reader. The vertical axis is a line starting with the text and moving down to the exterior texts, or contexts, of the text in question. This vertical axis emphasizes the text as a writing in relation with other writings. In other words, a text's dialogic interaction operates along both horizontal and vertical axes, from writing subject to addressee, from text to contexts. What emerges from a text is not "a point (a fìxed meaning)," but rather a dialogue of writer and reader, text and context (Desire 65).

In her early work, Kristeva's insistence on spatialization, which embeds her critique of pure formalism, is part of her Bakhtinian project to (re)insert the social and historical context as a necessary dimension of a text. Reading, she suggests, should never be merely a "linguistic" process focused on an isolated text. Consequently, she advocates a reading of what she calls the "translinguis­tic," by which she means the text's dialogue along horizontal and vertical axes with its writer, readers, and context (Desire 69). She coins the term "ideolo­geme" to identify the point of intersection between the text and its precursor texts. This ideologeme is "materialized" "along the entire length of its [the text's] trajectory, giving it its historical and social coordinates" (Desire 36). Reading for the dialogic ideologeme means reading the text "within (the text of) society and history" (Desire 37).

SPATIALIZING NARRATIVE

Kristeva's spatialization of the word has potential applications for narra­tive. I will alter her model of a text's vertical and horizontal axes at the same time as I maintain her insistence on historical and intertextual resonances. As an interpretive strategy (not as a narrative typology), I propose two kinds of narrative axes whose intersections are reconstructed by the reader in the inter­active process of reading. Bakhtin's notion of the novel's double chronotope is useful:

Even in the segmentation of a modem literary work, we sense the chrono­tope of the represented world as well as the chronotope of the readers and creators of the work . . . . [B]efore us are two events: the event that is narrated in the work and the event of narration itself (we ourselves partici­pate in the latter, as listeners or readers); these events take place in different times . . . and in different places.(255)

The totality of the work, he concludes, is made up of the interacting chrono­topes of the writer, reader, and text. We can graph Bakhtin's two chronotopes along horizontal and vertical narrative axes. The horizontal narrative axis in­volves the linear movement of the characters through the coordinates of textual space and time. The vertical narrative axis involves the space and time the writer and reader occupy as they inscribe and interpret what Kristeva calls the "subject-in-process" constituted through the "signifying practice" of the text and its dialogues with literary, social, and historical intertexts.

Both axes represent a movement through space and time—the one (hori­zontal) referring to the movement of characters within their fictional world; the other (vertical) referring to the "motions" of the writer and the reader in relation to each other and to the text's intertexts. Where the horizontal movement exists in finite form within the bounded world of the text, the vertical movement exists fluidly as a writing inscribed by the writer and reconstituted by the reader more or less consciously and to a greater and lesser degree depending on the specific writers and readers. As different functions of narrative, these axes feed off each other symbiotically; neither exists by itself as a fixed entity. I separate them only for strategic purposes, for the insight that such a spatialization provides for interpreting the overdetermined complexities of narrative. A fully spatialized reading of a given narrative text, as narrative, involves an interpretation of the continuous interplay between the horizontal and vertical narrative coordinates. The "plot" of intersection, "narrated" by the reader, is a "story" based on a reading of the different forms that intersection takes through time, that is, how the horizontal and vertical narratives converge and separate, echo and oppose, reinforce and undermine each other.

Let me specify in more detail what I mean by horizontal and vertical nar­rative coordinates. The horizontal narrative is the sequence of events, whether internal or extemal, that "happens" according to the ordering principles of the plot and narrative point of view. Setting, character, action, initiating "problem," progression, and closure are its familiar components—the focus of much tradi­tional narratology.See for example Barthes's "Introduction to the Structural Analysis of Narratives"; Genette; Chatman; Phelan; and Brooks's discussion of spatialization in Russian Formalism and French structuralism (Reading 16). The horizontal narrative follows and is constrained by the linearity of language—the sequence of the sentence that moves horizontally in alphabetic scripts is repeated in the horizontal movement of the plot from "be­ginning" to "end," however the categories of start and finish are customarily understood. Determined in part by historically specific narrative conventions, the forms of the horizontal narrative differ particularly in their handling of chro­nology, teleology, and narrative point of view—from the "well-made" to the picaresque or "plotless" plot; from the omniscient to the multiple, unreliable, or first-person narrator; from the epistolary to the embedded and complexly framed narratives. But for all forms, reading the horizontal narrative involves interpret­ing the sequence inscribed in the linearity of sentence and story. In simplest terms, we ask, who is the story about? What happens? Where? Why? What does it "mean"? Ulysses and Mrs. Dalloway, for example, plot the movements of their characters through the cities of Dublin and London on a single day in June. Reading the horizontal narrative axis, we focus on the exterior and interior actions and thoughts of Clarissa and Septimus, Stephen, Bloom, and Molly (as well as a host of others). As readers, we may imaginatively inhabit their space and time, to become what Peter Rabinowitz and James Phelan variously call the "narrative audience" that participates in "the mimetic illusion" (Phelan 5).See their distinctions between the "narrative audience" (which accepts the story as "real") and the "authorial audience" (which covertly remains "aware of the synthetic"-that is, con¬structed-nature of the narrative) (Phelan 5). Rabinowitz proposed the original distinction, which Phelan develops extensively in relation to his work on the rhetorics of character and progression. As with any text, the horizontal narrative is reconstituted in the process of reading. Its attendant meanings are consequently dependent on what Brooks calls the reader's "performance" of the text (37), what Ross Chambers refers to as the "performative force" of the narrative (4-5), and what Phelan identifies as "nar­rative dynamics." But in bringing the horizontal narrative to life, the reader (like the writer) nonetheless remains in a different space and time from that of the characters.

The vertical axis of narrative involves reading "down into" the text, as we move across it. The vertical does not exist at the level of sequential plot, but rather resides within, dependent on the horizontal narrative as the function that adds multiple resonances to the characters' movement through space and time. The palimpsest—a tablet that has been written on many times, with prior layers imperfectly erased—serves as an apt metaphor for the vertical dimension of narrative. Instead of the single textual surface of the horizontal narrative, the vertical narrative has many superimposed surfaces, layered and overwritten like the human psyche. Freud's image of the psyche as the "mystic writing-pad" serves equally well-for with this mechanism, the written impression remains embedded, but hidden, in the wax beneath the clean plastic siate ("A Note").

The point of these tropes is not to suggest a simple equation of the horizontal narrative with consciousness and the vertical narrative with the unconscious. Rather, they suggest that every horizontal narrative has an embedded vertical dimension that is more or less visible and that must be traced by the reader because it has no narrator of its own. Although not yet named as such, the vertical narrative has been the focus of much recent post-structuralist, feminist, Afro-Americanist, and Marxist narrative theory.See, for example, DuPlessis, Brooks, Gates, Bersani, Chambers, and de Lauretis.

Although interwoven, three distinct strands of the vertical narrative can be usefully separated for purposes of analysis: the literary; the historical; and the psychic. Both the literary and historical aspects of the vertical narrative involve reading the horizontal narrative's dialogues with other texts, interpreting, in other words, the various forms of intertextuality that Kristeva introduces in her tropes of spatialization. Whether consciously or unconsciously produced by the writer, these dialogues exist as "the mosaic of quotations" that traverse the text. They are the layered surfaces beneath and within the horizontal narrative, but they are not narrated by it and may seem tangential to it. When consciously intended by the writer, these intertextual resonances establish an indirect com­munication between writer and reader, with the characters and events of the horizontal narrative as points of mediation. Such resonances do not usually exist in the mind of the characters—in the space and time of the horizontal narrative. In Mrs. Dalloway, for example, it is the reader who "narrates" the story of Septimus as Shakespearean fool, as scapegoat, as sacrificial lamb and Christ figure within the anguished postwar landscape.

The literary aspect of the vertical narrative exists first of all in relation to genre. The writer's and reader's awareness of genre conventions exists as a chronotope, a space-time, within which the specific text is read—for its invo­cations and revocations, its uses and rescriptions, its repetitions and play. We read, for example, The Voyage Out, Woolf's first novel, within the grid of the Bildungsroman as a story of development that progresses conventionally through courtship and engagement, only to veer suddenly away from the mar­riage plot, in having its protagonist die at the end. More broadly, all literary texts exist—however centrally, ambivalently, or marginally—within one or more literary traditions or cultures. Horizontal narratives, consequently, have an indirectly narrated vertical dimension that accomplishes a dialogic engagement with what has been written before. In The Signifying Monkey, Henry Louis Gates proposes the term "signifyin(g)" to identify a culturally specific form of intertextuality, a mode tied to the African-American oral and written traditions of speakers and writers self-reflexively and intentionally playing off the dis­course of others in the tradition. The epistolary mode of Alice Walker's The Color Purple, in which Celie writes letters first to God and then to her sister, not only dialogues with such epistolary inscriptions of rape as Richardson 's Clarissa, but also signifies on the oral frame of Hurston's Their Eyes Were Watching God, in which Janie narrates the events of the story to her friend Phoebe, and on the story of incest in Toni Morrison's The Bluest Eye. More generally, we recognize that intertextual reference may be highlighted or muted, intentionally or unintentionally present, collaborative or revisionist. But com­mon to all intertextual resonances is a story of dialogue narrated by the reader that takes place outside the spatial and temporal coordinates through which the characters of the horizontal narrative move.

The historical aspect of the vertical narrative represents a similar mosaic of quotations, one that refers to the larger social order of the writer, text, and reader. Such a mosaic may involve reference to a specific historical event that the text reconstructs—such as Morrison's retelling in Beloved, with key depar­tures, of Margaret Garner's attempt to kill her children when faced with their and her own retum to slavery in 1856. Or, more broadly, this historical mosaic may involve what Rachel Blau DuPlessis calls "cultural scripts" layered into the horizontal narrative (Writing Beyond the Ending ix-xi, 1-19). DuPlessis' term acknowledges the part that "story" plays in both ideological and oppositional discourses. These political resonances that traverse the text might include inter­locking narratives of race, gender, class, ethnicity, sexuality, religion, and so forth—stories, in other words, that reproduce, subvert, and otherwise engage with the dominant and marginalized cultural scripts of the social order. For Fredric Jameson, such narratives constitute what he calls the "political uncon­scious," by which he means the "buried and repressed" narrative of class struggle present in trace form on the surface of the text (20). His assertion that narratives of class struggle subsume all other stories is dangerously bounded, but his call for the critic to read the text for signs of its repressed political scripts is useful. In Beloved, the vertically embedded cultural scripts or textual uncon­scious include many "stories" of race and gender relations: for example, the master's right to violate slave women; western theories of black and African inferiority and bestiality; pattems of slave resistance; white women's liminal position between race privilege and gendered alterity. Whether these politica! and historical narratives are buried in the text or openly scripted, reading this aspect of the vertical narrative allows for an analysis of the text in dialogue with "its historical and social coordinates," as Kristeva advocates (Desire 36).

Reading the psychic aspect of the vertical narrative involves recognizing that a text can be read as a linguistic entity structured like a psyche, with a conscious and an unconscious that interact psychodynamically. Freud's concept of the psyche as perpetually in the process of splitting suggests that nothing is ever lost, but only forgotten.This is, of course, a founding principle of psychoanalysis, made as early as Josef Breuer's and Freud's jointly written (Studies on Hysteria (1895). See also Freud's "Repression." Analogically speaking, the text is, like a dream, the result of a negotiation in which the desire to express and the need to repress force a compromise that takes the form of disguised speech. The text, then, can be read as a site of repression and insistent retum. Freud's grammar for the dream-work—the mechanisms of displacement, condensation, non-rational modes of representability, and secondary revision—is useful for decoding dis­guised expression (lnterpretation of Dreams 311-546). These mechanisms are often at work in the enigmatic textual sites that deconstruction unravels to sub­vert underlying binaries. As Shoshana Felman does in her reading of James's The Turn of the Screw, textual gaps, silences, knots, and aporias can be read vertically to gain some sort of access to the textual unconscious.

This is also, I would suggest, what Kristeva is doing in her integration of Lacanian psychoanalysis, Bakhtinian dialogics, and Barthesian semiotics in Revolution in Poetic Language.See especially Revolution (13-106), "From One ldentity to an Other" (Desire 124-47) and "Motherhood according to Bellini" ((Desire 237-70). For other formulations of the textual un-conscious, see Culler, Felman, Riffaterre, and Jameson. Kristeva inverts Lacan's axiom that the uncon­scious is structured like a language to suggest that the text is structured like a psyche. Language, she argues, always engages in a dialectical interplay of two modalities, the semiotic and the symbolic. The semiotic—that oral and rhyth­mic dimension of language that exists prior to and outside a system of signifi­cation—harkens back to the pre-oedipal period of the child's desire for the matemal body. The symbolic—that meaning-centered, instrumental aspect of language that exists after the child grasps the principle of signification—reverts back to the oedipal period when (according to Lacan) the child's realization of sexual difference allows for the acquisition of language based on a system of differences govemed by the Law of the Father ("Signification of the Phallus").See especially Kristeva, Revolution (13-106) and "From One Identity to an Other" (Desire 124-47). Reading for the interplay of the semiotic and the symbolic—newly and differ­ently constituted in every text—is one form that reading the vertical narrative can take.

A relational reading strategy based on the compositional history of the text—the chronotope of the writer—offers another mechanism for reading the psychic dimension of the vertical narrative. Instead of privileging the "final" text as the "definitive" one, we can read the various versions of a text as an overdetermined palimpsest in which each text forms a distinct, yet interrelated part of a larger composite "text."For a different attempt to move textual criticism beyond the teleologica} search for the "defin¬itive" text, see Jerome McGann's A Critique of Modem Textual Criticism . Freud's concept of dreams in a series as being part of a larger dream-text that can be interpreted is useful (lnterpretation of Dreams 369, 563). He suggests that the mechanisms of the dream-work govern the relation among the dreams and that "the first of these homologous dreams to occur is often the more distorted one and timid, while the succeeding one will be more confident and distinct" (369). Freud's grammar can be adapted to read what gets repressed on the one hand and worked through on the other in a series of drafts preceding a published text. Tracking the various versions of the same story can reveal a process of conscious or self-conscious self-censorship whereby textual revision of the horizontal narrative represses or further dis­guises certain forbidden elements that remain as part of the vertical narrative. Reading the composite "text" involves reconstructing the "story" of condensa­tion, displacement, and secondary revision from one version to another. In short, earlier versions of a text can be read vertically as the textual unconscious of the horizontal narrative in the published text. For example, H. D.'s "Madri­gal Cycle"—a triptych of three autobiographical novels about her life during the teens—forms a composite text in which the last one she wrote (Bid Me to Live [A Madrigal]) represses the stories of lesbian desire and illicit motherhood that are fully narrated in the earlier texts, Paint It To-Day and Asphodel, both of which she ultimately considered to be "drafts" for Bid Me to Live.I have made this argument more fully in "Return of the Repressed in Women's Narratives."

Conversely, a writer's repeated retum to the scene of writing a particular story can be read as a kind of repetition compulsion in which the earliest ver­sions are the most disguised, with each repetition bringing the writer closer to the repressed content that needs to be remembered. Here, I am adapting Freud's notion of analysis as a transference scene in which the analysand repeats the symptoms and dreams produced by repressed materiai in a process of "working through" that ultimately leads to conscious recollection of what has been for­gotten ("Further Recommendations"). This analogy between analytic and nov­elistic transference is especially cogent in autobiograpbical narratives, in which the split subject of the writing "I now" and the written about "I then" perform the different roles of analyst and analysand in a kind of "writing cure." Earlier or later versions of the horizontal narrative, in other words, can, when read together as a composite text, give us access to tbe psychic dimension of the vertical narrative. Joyce's autobiographical narratives about Stephen D(a)edalus, for example, (including "A Portrait of the Artist," Stephen Hero, A Portrait of the Artist as a Young Man, and Ulysses) constitute a composite "text" in which the death of Joyce's mother and the son's remorse remain unnarratable until the final text in the series—present only in trace forms that can be inter­preted with Freud's grammar for the dream-work.Like H. D.'s Madrigal Cycle, the different versions of the Dedalus narratives can be read both ways, with both early and late texts serving as the textual unconscious for the others in the series. See my "(Self )-Censorship and the Making of Joyce's Modernity" in Joyce: The Return of the Repressed. The essays in this collection provide examples of reading the vertical narrative axis in its literary, historical, and psychic dimensions.

WHY SPATIALIZE NARRATIVE?

What do we learn by conceptualizing narrative in spatial terms? In contrast to typological approaches, spatialization emphasizes the psychodynamic, inter­active, and situational nature of narrative processes; it also provides a fluid, relational approach that connects text and context, writer and reader. Spatiali­zation is not, of course, the only way to produce such readings. Other interpre­tative strategies have gained access to a text's literary and historical resonances without resort to spatial tropes. Other critics, such as Ross Chambers, have developed ways of reading what be calls "not the actual historicity of texts, but the markers, within them of historical situation" (10). His analysis of a text's contractual appeal and adaptation of Clifford Geertz's "thick description" rep­resents a different route to reaching some of the same objectives. James Phelan's distinction between the mimetic and synthetic dimensions of character as the reader experiences it represents yet another. For Phelan, the mimetic aspect suppresses the reader's awareness of the character as authorial construction while the synthetic foregrounds this construction as part of a communication between author and reader (1-27, 115). For him, it is the play between the mimetic and the synthetic that accounts for narrative progression, a theory that assumes, like spatialization of narrative axes, that the text operates on an inter­play between two different chronotopes—the mimetic world of the characters, and the synthetic realm of the author and reader.

Spatialization can, however, go beyond these other metbods by facilitating some new readings of narrative that might not otherwise exist. The notion of a vertical axis embedded in the horizontal suggests the way in which historical, literary, and psychic intertextualities constitute more than resonances attached to the text associatively, suggestively, or randomly. Instead, they initiate stories themselves—dialogic narratives "told" by the reader in collusion with a writer who inscribes them in the text consciously or unconsciously. This, I would argue, is the contribution made by Kristeva's graph of the writer/reader (hori­zontal axis) and text/context (vertical axis). Moreover, the concept of interactive horizontal and vertical narrative axes allows for a relational reading of the two that produces a "story" not present in either axis by itself. For example, the horizontal narrative of The Voyage Out is the story of failure, the Bildung that ends in death. The vertical narrative, reconstructed by the reader, is the story of rebellion, of Woolf's successful "voyage out" of the marriage plot, one that led out of the drawing room and into the world of letters, as an initial declaration of independence from the dominant literary and historical narratives of the early twentieth century. The confrontation in this case between the horizontal and vertical narratives constitutes a "story" of its own that is present in neither nar­rative axis by itself.

A "full" reading of narrative axes is not possible in a bounded text because, like the dream in Freud's psychoanalysis, the text's dialogism is unbounded, as is the story of the intersections between the horizontal and vertical coordinates. But a reading strategy based in the identification of horizontal and vertical nar­ratives axes fosters relational readings, discourages "definitive" and bounded interpretations, and encourages a notion of the text as a multiplicitous and dy­namic site of repression and return. Such spatialized readings also allow us as readers to construct a "story" of the fluidly interactive relationship between the surface and palimpsestic depths of a given text--taking into account all the historical, literary, and psychic resonances that are embedded within the horizontal narrative and waiting to become narrated in the reading process. ldeally such a story is made up of a sequence of relational readings that at every point in the horizontal narrative examines its vertical component. The rich­est insights produced by a spatialized reading strategy may well reside in the way it potentially produces interpretations of the textual and political un­conscious of a given text or series of texts. But in general, spatializing narrative gives us a systematic way of approaching the various forms of narrative dialogism and of (re)connecting the text with its writer and world. In Kristeva's words, spatialization suggests an interpretive strategy that regards a text as "a dynamic..intersection of textual surfaces rather than a point (a fixed meaning), as a dialogue among several writings: that of the writer, the addressee..., and the contemporary or earlier cultural context" (De­sire 65).

WORKS CITED