Beginnings and Endings as Textual Limina
Each temporal sequence (specifically in language) has its own structure and dynamics, but the beginning and the ending may be said to be universally important or significant points within such a sequence. We have plenty of evidence of this from the everyday use of language: for example, sayings about the crucial importance of “first impressions,”1 or expressions such as the “last word” and the “bottom line,” highlighting the importance of what comes at the end in determining the meaning of the whole utterance, or the fact that words referring to the termination point such as “conclusion” or “end” also have logical or teleological senses (a judgment reached by reasoning in the former case, a goal or destination in the latter). A consideration of beginnings and endings as distinctive units of a verbal sequence can already be found in the earliest studies that reflect systematically on the sequential organization of language (i.e., ancient Greek and then Roman rhetorical treatises).2
“Beginning” and “ending” are vast topics and, as terms, may be used in a wide array of meanings. This article’s focus, as the foregoing paragraph should have made clear, is textual, or discursive—first and foremost on the beginnings and endings of literary works (so that with regard to starting points, the term “openings” could have been less ambiguous). However, as the following discussion will illustrate, they may be interestingly related to, or closely intertwined with, beginnings and endings of other kinds or at other levels. Narrative texts, for example, are distinguished by the coexistence of two temporal sequences—the mimetic/actional in addition to the textual (in other words, that of the told as well as the telling/reading sequence)—and the interrelations between textual beginnings or endings and their mimetic counterparts are always worth looking into.3 Additionally, textual limina may be said to be particularly congenial to what Catherine Romagnolo calls “conceptual” explorations: namely, the thematic interrogation or theorization of beginnings and endings as well as related concepts (such as origins and destinies, respectively).4
One of the most famous pronouncements about the relation of an artwork to reality is by Henry James, in the New York edition preface to Roderick Hudson: “Really, universally, relations stop nowhere, and the exquisite problem of the artist is eternally but to draw, by a geometry of his own, the circle within which they may happily appear to do so.”5 The beginning and the ending may be said to play by definition an important role in the drawing of such a circle, by constituting the actual boundaries, or frame, of the literary text. These boundaries separate the text—and the world it projects—from the world around us, thus helping to determine its basic shape. The spatial metaphors of a circle or a frame, however, should not make us forget that the literary text is a temporal artwork and therefore that its beginning and ending, despite the symmetry they display as its “edges,” are also highly asymmetrical in terms of their logic within the temporal sequence. They typically perform different functions, operating within different sets of conventions and raising different kinds of issues.
The Beginning as a Threshold: Location and Functions
There may be a misleading simplicity in the very directness in which the term “beginning” is normally used with regard to literary works; locating the point of beginning is often somewhat complex or problematic—typically more so than that of the ending (an example of asymmetry). As a rule, readers do not approach the text “itself” (or the main text) directly, because at least since the advent of the print era and the book format, it is accompanied—or surrounded—by other materials that, following Gérard Genette’s seminal study, are known as paratexts (or, collectively, the paratext).6 Of the various paratexts discussed by Genette, those that may typically influence reception the most may be said to be the name of the author, title, epigraph, and prefatory materials (which may be variously titled—as “preface,” “introduction,” “prologue,” and so on).7 The paratext is likened by Genette to a threshold (rather than a clear-cut borderline) through which the reader gradually passes from the “outside” to the “inside” of a text.8 So by its very nature its status with respect to whether it is part of the text “itself” or not is ambiguous. Such a differentiation, moreover, tends to become particularly tricky when it comes to fictional paratexts—namely, those presented as composed by fictional characters rather than attributed (directly) to the author or any other nonfictional entity.
It is “well-known,” for example, that Herman Melville’s Moby-Dick (1851) opens with “Call me Ishmael.”9 However, that celebrated sentence—which begins the first chapter—is preceded by two sections of “Etymology” and “Extracts” (the latter quite lengthy), which may be attributed to the same narrator who invites us to call him Ishmael. Similarly, chapter 1 of Part 1 of Vladimir Nabokov’s Lolita (1955) begins with another well-known opening sentence: “Lolita, light of my life, fire of my loins”—preceded, however, by a less well-known foreword by “John Ray, Jr., Ph.D.” which begins “‘Lolita, or the Confession of a White Widowed Male,’ such were the two titles under which the writer of the present note received the strange pages it perambulates.”10 And whereas in both these cases the beginning that has established itself in the collective cultural memory is that of the “main” text rather than of a prefatory paratext, such is not necessarily always the case. Another famous opening sentence, that of Ralph Ellison’s Invisible Man (1952), is “I am an invisible man”—the beginning of the prologue; the novel’s chapter 1 begins: “It goes a long way back, some twenty years.”11 Thus, the perception of what constitutes the “beginning” may be influenced by sheer memorability (or quotability) rather than any clear formal criteria.
Of course, authors may be fully aware of this measure of ambiguity or fuzziness with regard to the point of beginning and take advantage of it—as by devising what amounts to two (or even more) beginnings, each with its own functions. A sophisticated example of this approach may be found as early as the 16th century in the anonymous Life of Lazarillo de Tormes (1554), the founding text of the picaresque genre. There, chapter 1 opens in chronological order with the narrator-protagonist’s origins and birth, slowly building an arc of suspense concerning his early life, whereas the prologue that precedes it ends by producing curiosity about the narrator’s situation and motives for telling the story of his life at the very moment of writing, when he addresses a certain “Your Honor,” mentioning some “matter” or “case” (caso in the original Spanish) this addressee has asked him to relate in detail.
If the beginning is a threshold, then one of its most important potential functions is naturally to “draw us in,” or to be seductive—help carry us over from the world we inhabit to the world the author has imagined. Thus, though beginnings may certainly be low key, it is no coincidence that a relatively high number of memorable literary passages (particularly at the level of the single sentence) are openings; since there is nothing previous to “support” it at this stage, the text needs to manifest a high intensity in order to be enticing.12 James Phelan enumerates four main “means of enticement” employed by beginnings: (1) initiating us into a striking relationship with the two main communicators—author and narrator (who may converge or diverge); (2) plunging us into an unstable situation; (3) introducing us to a remarkable character; and (4) striking a significant and compelling thematic note (think of how the beginning of Jane Austen’s Pride and Prejudice [1813] introduces the theme of marriage, or that of Leo Tolstoy’s Anna Karenina [1878] the theme of the family).13 A fifth type of enticement can be added: introducing a rich and evocative setting (think of the detailed descriptions of houses in various Balzacian novels, the description of London in the fog at the beginning of Charles Dickens’s Bleak House [1853], or that of Egdon Heath, which opens Thomas Hardy’s The Return of the Native [1878]).14
As Phelan also points out, these categories are definitely not mutually exclusive. For example, the well-known opening sentences of the three American novels quoted earlier combine (1) and (3) in several ways, most basically because in hom*odiegetic (“first-person”) narratives the narrator is perceived as a character within the storyworld, so that the act of narration also contributes to his—or her—characterization. Or consider the opening sentence of Austen’s Pride and Prejudice, already mentioned as striking a significant thematic note—“It is a truth universally acknowledged, that a single man in possession of a good fortune, must be in want of a wife.”15 It is also notable in terms of striking up a relationship with a witty and observant narrator (in this case, a representative of the author) who establishes herself as the reader’s “friend and guide” (to use a phrase employed by Wayne Booth in relation to another novel by Austen, Emma).16 This is done in a manner that requires minute attention to the ironic inflections of her voice and the perspectival manipulations it performs. The gnomic and apparently authoritative tone of the generalization about the “single man in possession of a good fortune” is immediately undercut by a non sequitur (why must such a man be in want of a wife?); this non sequitur is revealed in the next sentence as representing a certain subjective and biased perspective within the storyworld—the collective viewpoint of “the minds of the surrounding families” in the neighborhood this single man is entering, who consider him as “the rightful property of some one or other of their daughters,” however little “is known [of his] feelings or views.”17
The contemporary German author Martin Walser wrote: “The novel is always a game, but the first sentence inexorably determines how to play it.”18 This is clearly an exaggeration, with regard both to the shortness of the sentence as a unit sufficient for such a determination and to the inexorability of the latter. But the exaggeration does suggest a valid point: namely, that the opening—as that stage whose reception is not conditioned by any previous stage of the text continuum—is particularly effective in setting off our mind in a certain direction, influencing our entire reception of the work. Thus, it may play a particularly important role in negotiating what is sometimes called the “contract” between author and reader; such a role is exemplified by how innovative works often begin with a pronounced display of (at least some of) their innovative aspects.
Consider the opening paragraph of Daniel Defoe’s Robinson Crusoe (1719), in which the narrator-protagonist provides us with chronological, geographical, and genealogical details relating to his birth and family (going on to provide an etymology of his name).19 This amassing of “objective” data in a dry tone, lacking psychological or emotional intrigue, may seem quite unremarkable but in its historical context is noteworthy. It signals the realistic revolution of the novel in the 18th century as described by Ian Watt in The Rise of the Novel—by foregrounding the referential aspect of language, emphasizing the individuation of character with specifics of time and place or the use of a full name (of the kind familiar to readers from the contemporary social environment), and already giving a clear indication of how, instead of relying on traditional plots, the plot of the novel would be subordinated to the pattern of the autobiographical memoir (in “as defiant an assertion of the primacy of individual experience [over collective tradition] . . . as Descartes’s cogito ergo sum was in philosophy”).20
Nearly two centuries later, James Joyce’s A Portrait of the Artist as a Young Man (1916) played an important role in establishing the modernist poetics of the novel, featuring what may be viewed as a new kind of realism with a much greater subjective bent. Thus, the opening of this novel dispenses with orderly narration (or expositional presentation), providing us instead with a series of (highly sensory) impressions in an attempt to capture the perceptions of a young boy. In addition to the use of childlike speech and syntax, this is done through the narrative syntax itself—the jumps from impression to impression with no clear overarching chronological sense reflecting (and motivated by) a childlike consciousness, or the fragmentary memories one typically has of this early period of life.
Orientational Versus Abrupt Beginnings
A detailed history of literary beginnings is clearly well beyond the scope of this article, but I would like to focus on one important development indicative of wider trends. Consider the following beginnings of Jane Austen’s Emma (1816), and Ernest Hemingway’s “The Short Happy Life of Francis Macomber” (1936), respectively:
Emma Woodhouse, handsome, clever and rich, with a comfortable home and happy disposition, seemed to unite some of the best blessings of existence; and had lived nearly twenty-one years in the world with very little to distress or vex her.
. . .
It was now lunchtime and they were all sitting under the double green fly of the dining tent pretending that nothing had happened.21
These may be broadly considered as representative of two types: “traditional” and “modern.” The latter, which appears in Hemingway’s story, would have been almost unimaginable before the late 19th century. In the example of the former type, the narrator names the protagonist and goes on to provide important information about her, summarizing the first twenty-one years of her life in order to present them in overview; this narratorial activity strongly signals an acknowledgment of the need to guide and orient the reader while introducing him (or her) to an unfamiliar world. In the example of the latter type, by contrast, a sense of abruptness prevails as the reader is directly confronted with an ongoing action without being supplied preliminary information necessary for its understanding (as if it is assumed the reader is already familiar with things that are in fact still unfamiliar).
Several typical devices employed by Hemingway to achieve this effect may be noted. First, the obscure (in terms of its reference) pronoun “they.” This is an example of what Joseph Backus, in an important article that described this phenomenon based on a large corpus of American short stories, called “a nonsequential sequence-signal”—namely, a word that normally depends for its full meaning on previous information but appears without this information.22 And as a third-person pronoun, the plural “they” has an even stronger effect than a single “he” or “she,” since it is not known how many people it refers to.23 Second, the immediate use of definite articles in relation to objects—as though there already exists a familiarity with these objects, turning the noun phrases that follow the articles into nonsequential sequence signals as well (“the double green fly of the dining tent”—contrast with “a comfortable home and happy disposition” from the beginning of Emma). And third, there is a clear implication of some anterior, significant action, affecting the present: something that the characters pretend did not happen (What was it? Why do they feel the need to pretend?).24
Backus identifies Henry James as the writer who established as a regular device the use of nonsequential sequence signals such as referent-less pronouns and familiarizing definite articles in the beginning of stories; Georg Bossong, in a parallel study based on a corpus of French literature, traces this tendency to Guy de Maupassant and Émile Zola.25 In both cases the rise of the new type of beginning dates back to the late 19th century and seems to coincide, more generally, with the transition from realism to modernism. Specifically, the modernist tenet of “Exit Author”—erasing the visible and obtrusive authorial persona as represented by a narrator-figure—is well served by a beginning that backgrounds the presence of such a narrator (rather than foregrounding the narrator as the reader’s guide to the fictional world), thus minimizing (or at least moderating) the sense of mediation.26 Franz Stanzel, in his discussion of different types of narrative situations, considers beginnings of the “abrupt” kind as prototypical of a reflector-character (as opposed to teller-character) situation, in which the story is transmitted to the reader predominantly from the perspective of a non-narrating character—a narrative situation that indeed became predominant in modernist fiction.27 The absence of narrative preliminaries and the use of devices such as nonsequential sequence-signals are, Stanzel claims, particularly effective as means by which “reflectorization” of the narrative situation can be achieved immediately, with the reader being placed in the position of the reflector-character, experiencing the narrated events in actu. (The reflector-character is typically the one to whom the initial “referentless” pronouns refer; in Hemingway’s story, all three characters to whom the initial “they” turns out to refer are used as reflectors to one extent or another.) At any rate, it should be noted that regardless of the historical (and poetic) context in which beginnings of this kind developed, they now constitute an important part of the narrative repertoire at the disposal of any writer.
Abrupt beginnings naturally raise the notion of in medias res (“in the midst of things”), a Latin term originating in Horace’s Ars Poetica (c. 15 bc), where he praises Homer for beginning the story of the Trojan war in this manner rather than ab ovo (“from the egg”—i.e., from Helen’s birth). Since the Renaissance, this statement has usually been interpreted as a recommendation, addressed to epic poets, to begin their works by plunging into the middle of the action, and the term “in medias res” has come to denote in general this kind of deformation of chronological sequence—beginning the textual sequence from somewhere in the middle of the mimetic sequence of the action; or, in the terminology of the Russian formalists, the sujet in the middle of the fabula.28 But it is important to note that such a deformation in itself should be differentiated from the “perspectival abruptness” of the non-orientational type of modern openings.29 The latter certainly lends itself readily to chronological in medias res deformations (as the opening of Hemingway’s story demonstrates) and tends to produce a particularly strong in medias res effect because of the foregrounded presuppositions of missing information. However, not every chronological in medias res beginning necessarily displays such perspectival abruptness. Consider the opening of the text that came to be considered as the quintessential example of in medias res, Homer’s Odyssey:
Tell me, Muse, the story of that resourceful man who was driven to wander far and wide after he had sacked the holy citadel of Troy. He saw the cities of many people and he learnt their ways.30
This is certainly in line with the “traditional” type of beginning. The presence of the narrator as a mediator of the storyworld becomes pronounced already with the initial invocation of the muse and then further with the introduction of the protagonist. This introduction, however, is far from exhaustive and clearly designed to arouse interest about Odysseus and his adventures since the end of the Trojan war—in which context the short glimpses provided into these adventures in the following sentences, as well as the beginning of the ongoing action at the relatively late stage of Odysseus’s stay on Calypso’s island, come into play.
The concept of in medias res raises the more general issue of the interrelations between the mimetic and textual sequences of a narrative, and with respect to beginnings a particularly interesting and important aspect of this interrelationship has to do with phenomenon of the exposition.
Exposition, Fictive Present, and the First Discriminated Occasion
The exposition is particularly intriguing in the context of narrative beginnings since by definition it always constitutes the beginning of the mimetic sequence: namely, the chronologically ordered sequence of motifs as reconstructed by the reader. But it is not necessarily located at the beginning of the textual sequence, and even if it appears there it is not necessarily limited to it but rather can be distributed throughout the work. Moreover, the point of transition between the exposition and the primary narrative action may be considered as another kind of “beginning,” which plays an important role in how the narrative is perceived as a whole.
In the major study of the exposition to date, Meir Sternberg’s Expositional Modes and Temporal Ordering in Fiction, its function is described as “[introducing] the reader into an unfamiliar world, the fictive world of the story, by providing him with the general and specific antecedents indispensable to the understanding of what happens in it.”31 And from the examples Sternberg immediately goes on to provide for such antecedents, which include the traits and habitual behavior of the dramatis personae and the canons of probability operating in the storyworld, he is clearly thinking not only of prior “events” in the common meaning of the term, but also of the general “state of affairs” at the starting point of what is perceived as the primary narrative action—to which, following Sternberg, this article will refer as the fictive present. The relation between the exposition and the fictive present may be compared to that between “background” and “foreground” according to gestalt perception theory.32 This is with the understanding that being part of the background does not entail unimportance or insignificance by any means: consider, for example, the standard detective plot in which the whole story of the investigation in the fictive present often revolves around the attempt to uncover expositional information.
So how is the exposition, as the beginning or the first part of the fabula, delimited? Its terminus a quo would consist of the chronologically earliest motifs in the narrative and its terminus ad quem of where the fictive present begins; Sternberg tackles at length the question of how to determine the latter, namely, the temporal point in the fabula that marks the end of the exposition.33 A key concept introduced for this purpose is that of the first discriminated occasion. The idea of a discriminated occasion has to do with scenic presentation: Every work establishes a certain scenic norm of its own, in which the duration of the (textual) representational time approximates that of the (mimetic) represented time. Such scenic discrimination generally implies greater centrality for the materials (or time span) treated this way compared with those that are briefly summarized or skipped altogether. More specifically, the first discriminated occasion in the textual sequence assumes a special conspicuousness and significance, the author finding it to be a time section of enough consequence to deserve the first full scenic treatment—implying, according to Sternberg, that this is what he or she has decided to make the reader regard as the beginning of the fictive present. And once this temporal signpost is established, the reader should be able to determine what motifs are expositional in relation to it—namely, those anterior to it in the fabulaic order—no matter what their position in the sujet.
When the work begins with a full-fledged scene, as in William Thackeray’s Vanity Fair (1848) or James’s The Ambassadors (1903), it signals right away the point where the fictive present begins and the exposition ends.34 Things may be less obvious if the work begins with a preliminary and concentrated exposition—a continuous block of antecedents. Sternberg enumerates a set of criteria that allows the reader to recognize where the “action,” or fictive present, starts in such a case; its scenic texture would distinguish itself from that of the exposition by its specificity (degree of detail with regard to whatever exists or takes place in the course of the represented time), concreteness (uniqueness in time and space), and actional dynamics (a disturbance of a once-stable state of affairs instead of the description of such a state). Two examples of these criteria at work are the scene with the dialogue between the Lord and Satan narrated in verses six to twelve of the Book of Job (following the preliminary exposition in the first five verses), and the scene that takes place in the first chapter of Austen’s Emma after Miss Taylor’s wedding, with Emma Woodhouse, her father, and later Mr. Knightly (following a preliminary exposition that telescopes into a couple of pages the first twenty-one years of Emma’s life). In addition to such relatively straightforward cases, Sternberg notes potentially more tricky ones of what may be called pseudo-scenes (or pseudo-discriminated occasions) within the preliminary exposition—namely, scenes that perform an illustrative function and are thereby deconcretized and deactionalized. An example can be found in Edgar Allan Poe’s “The Murders in the Rue Morgue” (1841), the founding text of the detective genre.35 It begins with a clearly expositional discourse—some general reflections on the topic of the analytic mind, followed by a sketch of C. Auguste Dupin and a description of how he and the nameless narrator lived together in Paris, a textual segment composed almost entirely of essentially static and recurrent motifs (except for the short mention of the duo’s first meeting). It is then followed by what might be understood at first glance as a discriminated occasion, in which Dupin performs a feat of mind reading while he and the narrator are strolling the city at night. However, within the overall narrative framework this scene has no direct sequel and, moreover, is explicitly introduced as an example. (“But of the character of his [Dupin’s] remarks at the periods in question an example will best convey the idea.”36) Therefore, it is best integrated into the tale’s structure of meaning as an expositional illustration of Dupin’s superior reasoning powers. The true beginning of the fictive present (and the detection plot) comes only in the next scene, with the narrator and Dupin reading about the extraordinary happenings at the Rue Morgue.37
Sternberg provides throughout his study many examples of the first discriminated occasion demarcating the beginning of the fictive present from the exposition, and a huge number of additional examples may be found on reflection. This indicates an interesting tendency to a significant “synchronization” between the sequences of fabula and sujet even in cases of in medias res openings.
However, as long as we do not accept that the first discriminated occasion determines the beginning of the fictive present just by definition but rather want to examine if this is indeed the case by considering how the motifs that are anterior to the first discriminated occasion in the fabulaic order are treated in the work, we would be able to find counterexamples as well—that is, of first discriminated occasions that cannot be reasonably said to mark the beginning of the fictive present. Perhaps the clearest cases of this kind may be found in works beginning at the end—or near the end—of the chronological sequence of events. Consider Tolstoy’s novella The Death of Ivan Ilyich (1886), which begins with a short scene taking place in the St. Petersburg law courts, where three friends and colleagues of Ivan Ilyich learn from the newspaper of his death; there follows in the first chapter a rather lengthy scenic sequence in which one of the three, Peter Ivanovich, goes to the wake at Ivan’s house that night out of a sense of obligation. Next, the second chapter circles back to the beginning of Ivan’s life and commences with his biography. Thus, the whole story of Ivan Ilyich’s life, including the detailed and dramatic narration of his dying, is anterior in the fabulaic order to the first discriminated occasion (which takes place after his death), but it would be rather absurd to consider all of it as exposition. Somewhat similarly, in Malcolm Lowry’s Under the Volcano (1947) the first chapter also constitutes an “epilogue” in terms of the fabulaic order of events, in this case taking place a full year after the fictive present of the novel’s remaining chapters, which follow the final day of the protagonist’s life.38 One can also mention in this context works that begin at a late stage of the fabula but then do not circle back to the beginning of the action right after the opening but rather progress gradually from later to earlier events, such as Elizabeth Jane Howard’s The Long View (1956) or Sarah Waters’s Night Watch (2006).39 A major effect of such retrograde narratives may be said to be precisely the blurring of the borderline between exposition and fictive present, since motifs that can be perceived as expositional in relation to later events keep moving into the foreground of the narration.
In sum, both the exposition and the point of transition to the fictive present are important “beginnings” on the level of the mimetic sequence of narrative texts, and both meaningfully interrelate with the level of the textual sequence. Within this context, the first discriminated occasion in the textual sequence often signals the said point of transition and thereby a significant synchronization between the two sequences is produced; but the first discriminated occasion is not bound to perform this function. Even when it does not, its choice—namely, the choice of the time section, which is of enough consequence to deserve the first full scenic treatment—may still be assumed to be significant on aesthetic and rhetorical grounds within the narrative whole.
The Ending as a Privileged Position in the Textual Sequence
We have already touched on the issues involved in delimiting the beginning as a textual unit. Similar problems typically exist with regard to the ending—at least as to where it begins; but in this case there is a further fundamental issue to note. Since the ending follows everything else in the text, it is quite difficult (if at all desirable, at times) to consider it without considering through it, so to speak, the text as a whole. Indeed, in addition to the sense of the final textual unit, the term “ending” is quite often used in a somewhat different (but closely related) sense of an Archimedean point arrived at when the text has been read in full, from which it can be viewed as a totality.
A suggestive testimony to how much an appreciation of the ending depends on what precedes it may be found in comparing two lists, of a hundred best first and last lines from novels, constructed by the journal American Book Review based on surveys among respondents active in the literary field (in 2006 and 2008, respectively).40 Relative to the opening lines, it is striking how many of the final lines chosen are unremarkable in themselves, without knowledge of the full works to which they belong. For example, one of the top-ranked last lines (no. 7) is “He loved Big Brother” from George Orwell’s Nineteen Eighty-Four (1949). A certain measure of the particularly chilling effect of this simple sentence may survive for non-readers through the meaning acquired by the expression “Big Brother” due to Orwell’s novel. But a real appreciation requires an understanding of how effectively the sentence concludes the novel’s full plot, by signaling the end of the protagonist’s rebellion and the extent of the party’s ability to control the individual—not just in terms of behavior and thoughts but even of emotions (and with that, investing the narrative’s warning about totalitarianism with particular force).
Clearly, there is a dialectic at work here. The understanding and appreciation of endings depend to a large extent on what has preceded them, but at the same time they tend to play an important role in retrospectively shaping it and to have a lasting impact on its evaluation. In general, we expect endings, more than any other part of a work, to show what the text was “about,” what it was trying to achieve (let us again recall the teleological sense of the word “end”). As the point beyond which nothing further follows, the ending becomes the text’s final opportunity of influencing the reader—which leads us to assume that, as Marianna Torgovnick puts it, “An ending is the single place where an author most pressingly desires to make his points—whether those points are aesthetic, moral, social, political, epistemological, or even the determination not to make any point at all.”41
Somewhat paradoxically, a testimony to this special status of the ending within the totality of the work is provided by the relative “instability” of endings, namely, the numerous documented cases of alterations that were made in them—more than in other parts of the text—especially when it comes to substantial changes in the narrated world itself (beyond the level of stylistic revision).42 This indicates that altering the ending can strongly influence the work as a whole even without taking the trouble of meticulously rethinking and rewriting that whole: a substantial effect achieved with relatively little effort. The best-known literary case is probably that of the alternate endings of Dickens’s Great Expectations (1861), and it is typical in at least two respects. First, it was the result of what might be termed post-completion pressures—the original ending met with some resistance or negative response that led to its modification; here the modification occurred in response to the suggestion of Edward Bulwer-Lytton, whom Dickens apparently considered a reliable representative of the reading public’s preferences. Second, the direction of the revision was from a less conventional ending by period standards (in which the novel’s protagonist is brought neither to the marriage altar nor to the grave) to a more conventional one (in which he sees no shadow of another parting from Estella).43 Therefore, such cases of altered endings may provide us with insight as to the range of conventions that operate in the relevant literary context, by highlighting the conventions whose undermining is perceived as disturbing in the original ending.
Ending and Closure
The critical study of the ending has paid a particularly great deal of attention to closure, so much so that there is a widespread tendency to conflate the two concepts; it is important, however, to differentiate between them. Whereas ending refers to the text’s termination point, which is an inevitable (and hence “obvious”) phenomenon in the sense that every text has to end somewhere, closure refers to the sense of an ending (to invoke the title of Frank Kermode’s well-known study).44 That is, closure is not the textual termination point itself but rather a certain effect, or perceptual quality, produced by the text; in Barbara Herrnstein Smith’s formulation, “one of stable conclusiveness, finality, or ‘clinch.’”45 This is certainly not an inevitable phenomenon but rather one whose creation may require complex and sophisticated strategies. Some texts may fail to create it or intentionally refrain from creating it; and in this case, we would be referring to “openness” (or open-endedness)—a polar concept that belongs to the same metaphor or semantic field. This distinctness of ending and closure does not mean, however, that they are not strongly related. As the text’s termination point, especially if considered as intentionally designed as such, the ending obviously plays a crucial role with regard to our ability to determine the nature and degree of the text’s closure. As long as the end has not been reached, such judgments would always be open to readjustment (including a complete reversal), because what is left of the text may contain further developments: what appears open may yet be closed, and what appears closed may yet be reopened.
Closure’s appeal to contemporary criticism as a topic probably has both a universal aspect and a more specifically historical one. Universally, it is a fundamental aesthetic and cognitive issue, whose consequences overlap with several other fundamental issues such as order, unity, and completeness. It seems to have a powerful existential dimension as well—consider how “closure” is often mentioned as important in evaluating various life experiences (such as relationships). From a more specifically historical perspective, problems and questions related to closure have been foregrounded by the downgrading of strong closure alongside the valorization of openness—in meaning generally and endings in particular—in modern(ist) literature as well as criticism.
Barbara Herrnstein Smith’s book Poetic Closure was a milestone in the study of closure, making a great contribution to the concrete understanding of how it is produced by literary texts. Smith considers closure as an effect produced by the text, a rhetorical or communicative goal (though she herself does not use these terms), and analyzes many textual devices employed as means for achieving this goal. Following Smith’s study, which focuses on poetry, attention in this line of investigation shifted to narrative, and the most detailed later studies of closural devices were done in German criticism, by Barbara Korte and Constanze Krings.46
A detailed survey of closural devices would be beyond the scope of this article; but since its topic is beginnings and endings, it is worth elaborating on one such basic and common device that involves the creation of strong linkages between them. Two spatial metaphors (already mentioned in the introduction) may be relevant for explaining the effect of closure produced by this device: those of the circle and of the frame. Creating the feeling that the ending has returned in some sense to the original point of departure, as in the drawing of a circle, suggests that there is no place else to go and thereby weakens the expectation of continuation.47 Alternatively, the linkage between beginning and ending may be perceived as drawing a frame around the work, with the suggestion that the ending, by completing the frame, is indeed the appropriate point to finish the work—thereby, again, weakening (or even canceling altogether) the expectation of a continuation.
Smith distinguishes in her study between structural and nonstructural (or local) strategies of closure. The former strategies relate to, or depend upon, the work’s overall structural principles (as in various plot resolutions, on which more in the next section, “Narrative Closure and Open(ended-)ness”); the latter are terminal features that tend to have closural force more or less independently of the work’s particular structure. An example of such a nonstructural strategy would be what Smith terms “closural allusion.”48 This is a reference at the end to termination, repose, stability and the like, or to events which in our nonliterary experience are associated with such qualities (e.g., death, sleep, night, or departure). Strong linkages between beginning and ending may be said to occupy an interestingly intermediate position between the two types of strategies: by definition, they produce at least a minimal “structure” by connecting two points in the text, and they may be tied to various other elements along the textual sequence—though this, in turn, might blur the distinctiveness of the framing.
The linkage between beginning and ending may produce its effect almost regardless of what comes in the middle, especially in relatively short texts, such as the following poem by an eight-year-old quoted by Smith:
If I were an Indian girlI would have a reddish-brown faceAnd not have to wear any lace.I would live in a pretty wigwamAnd wear a belt of wampum.I would walk the path that was narrowAnd carry a bow and arrow—If I were an Indian girl.
The basic dynamics generating this poem is that of the list, or catalogue, made of details about what the life of an “Indian girl” would be like; but such a dynamics does not yield in itself a termination point since, in principle, the list can go on and on. Thus, “having exhausted her impulse or imagination after seven lines, wanting a finished poem and knowing that she did not have one yet, the poet artfully repeated the first line—and was probably no less surprised than delighted to see how effectively it turned the trick.”49
Repetitions that connect the beginning with the ending do not have to be verbal but can also appear at the level of the represented world (and those of the former type often point to those of the latter), as in Conan Doyle’s The Sign of the Four (1890), which begins and ends with Sherlock Holmes injecting himself with cocaine to fight boredom. Moreover, symmetries between the beginning and the ending do not have to consist of exact repetition; it is enough that the reader perceives some kind of a significant equivalence pattern. For example, the Iliad begins with an old man (Chryses) coming from Troy to the Greek camp to ransom a captive child (his daughter Chryseis); it ends with another old man (Priam) coming from Troy to the Greek camp to ransom a captive child (in this case, the body of his dead son Hector). In addition to the difference stemming from the fact that in the later occurrence the goal is to ransom a dead body, the pattern involves a sharp contrast in that the earlier request is brutally denied (by Agamemnon), whereas the later one is granted (by Achilles).50 On an even higher level of abstraction, one can recognize at both the beginning and the ending of Austen’s Emma love triangles based on misunderstandings involving the protagonist—in both cases she mistakenly believes that a certain man (Mr. Elton at the beginning, Mr. Knightley at the end) is in love with and wants to marry Harriet Smith, her protégé, whereas in fact the object of this man’s desire turns out to be Emma herself. Here, indeed, the meaning of the term “beginning” is already stretched quite a bit (though it is still legitimate to use it in the context of this particular pattern), since the affair involving Elton occupies approximately the whole first quarter of the novel.
Such analogical linkages, in addition to producing the closural effect itself, may often acquire a special thematic or interpretative significance, since the juxtaposition of the beginning with the ending summarizes for the reader the full distance traversed between these two points or stages as if in a single “flash,” perceived supra-sequentially.51 Inasmuch as similarity predominates in this analogical relation, it foregrounds stasis and the lack of substantial change (as in Holmes’s unchanging situation and character), whereas if the contrastive aspect of the analogy is predominant, it foregrounds elements of change or development (the grimmer context of the later ransom request in the Iliad, alongside the more humane response it elicits; the great differences in Emma’s behavior, as well as its consequences, between the two affairs in Emma).52
The linkages between beginning and ending may also be more directly narrative (rather than analogical). For example, the meaning of the work’s title may be obscure and function like a mystery solved only at the end; one of the best-known uses of such a device is found in Stendhal’s Charterhouse of Parma (1839), where the title is revealed only on the final page as referring to the last place where Fabrice, the novel’s protagonist, withdraws before dying.53 The beginning and the ending may also form (or participate in) a mini-plot that foregrounds the narrative’s mediation. Such is the case in Thackeray’s Vanity Fair, where a preface titled “Before the Curtain” and attributed to the “Manager of the Performance” discusses the novel’s world about to be unraveled in terms of a play about to be performed by a puppet theater, whereas the novel’s final words announce the end of this play: “Come, children, let us shut up the box and the puppets, for our play is played out.”54 (Various types of frame stories may be viewed as elaborations of this device.) The last couple of examples also illustrates a point made earlier about the status of the paratext: When linking the ending to the beginning, this “beginning” may consist for the authors in paratextual elements (title, preface).
Narrative Closure and Open(ended-)ness
One of the most productive aspects of Smith’s approach to closure is her emphasis on the need to view it in relation to the dynamics of the text as a whole. As she notes at the beginning of her book: “It will be useful to regard the structure of a poem as consisting of the principles by which it is generated or according to which one element follows another. The description of a poem’s structure, then, becomes the answer to the question, ‘What keeps it going?’ . . . It allows the possibility of a corollary question, namely, ‘What stops it from going?’ and immediately suggests the close relationship between poetic structure and closure.”55 However, since Smith focuses on poetry (and mostly short lyric poems) she deals very little with narrative structures and the problems specific to them. In order to better understand narrative closure, Smith’s general approach can be synthesized with Meir Sternberg’s approach to the definition and understanding of narrative—and in particular of “narrativity,” namely, what constitutes the essence of narrative.56
Unlike most narratological approaches (since Aristotle), Sternberg’s defines this essence not in the mimetic terms of represented or narrated action but rather in rhetorical-communicative terms of narrative interest. This narrative interest is aroused in the reader through informational gaps regarding any aspect of the represented storyworld—be it an event, a motive for action, a character trait, a relationship, a viewpoint, a picture of society, or even an entire reality model. Most basically, such gaps result from the interplay between narrative’s two basic temporalities, which we have already discussed—the actional and the textual. Sternberg differentiates among three fundamental types of narrative interest according to a combination of two basic criteria: (1) Does the information withheld from the reader, via gaps, belong to the narrative past or future (both measured relative to what is perceived at any given point along the textual sequence as the narrative “present”)? (2) Is the reader aware that information is being withheld? In other words, do we know that we do not know? An awareness of not knowing would create an expectation for receiving the missing information: This would result either in suspense, if the expectation relates to the narrative future, or in curiosity, if it bears on the narrative past. However, if the reader is unaware that an informational gap exists, this gap will become perceptible and operative only when unexpectedly disclosed, thereby creating surprise.
Since Sternberg’s concept of narrativity is defined in communicative-rhetorical terms, it can be quite readily connected with Smith’s approach to closure. If, according to the latter, the key question one needs to answer in order to describe a text’s structure is “what keeps it going?,” then the former may be said to supply the fundamental answer with regard to a narrative text: namely, that narrative interest, in its three master types, is what keeps narrative going. The complementary question posed by Smith (“What stops it from going?”) directs us to consider how closure is created (or not), and in terms of Sternberg’s model the answer to that would be: the cessation, or termination, of narrative interest (in other words, the filling in of all the significant informational gaps about the represented world that have arisen along the textual sequence) is what stops narrative from going. Conversely, narrative “openness” would be the result of significant gaps relating to the represented world that remain open—or at least not definitely closed—even at the end; in other words, it would result from permanent gaps. It should be emphasized that, strictly speaking, what has been defined here is not the (non)closure of narrative but rather that of narrativity. That is, an explanation in terms of (non)termination of narrative interest does not necessarily relate to the (non)closure of everything within a narrative text; for example, equivalence patterns connecting beginning and ending may have nothing special to do with narrative in themselves. However, inasmuch as narrativity is dominant in the text, an explanation of closure in terms of manipulations of narrative interest would always have a crucial importance. And, moreover, the understanding of how any other elements in the narrative text operate for (or against) closure will depend on their interaction with these workings of narrative interest.57
This definition of the conditions for narrative closure or openness should clarify, among other things, that the distinction between “closed” and “open” endings is crude in its basic form and should not be regarded as (or reduced to) a simple dichotomy but rather be perceived as a finely gradated and multidimensional continuum. Any attempt to indicate the placement of a narrative text along this continuum should take into account the varied aspects of narrativity—such as different lines of narrative interest developed in a text, their different natures, their interrelations and relative hierarchies, the measure in which each of them reaches closure, and the combined effect of all these factors.
Let me provide a concrete demonstration of how the model presented above might work, with the detective genre as an example.58 For my purposes, it has the double illustrative merit of having a relatively clear-cut plot structure and of being a paradigm case for strong closure. The detective plot centers in the emergence of a mystery (typically related to a crime), leading to a systematic investigation aimed at solving it, usually conducted by a detective figure. The fictive present of the detective story moves, as a rule, along the axis of the story of investigation, which aims at solving the crime mystery; and because the goal of the investigation is, naturally, to fill in gaps related to past events, the dominant kind of interest generated by the detective plot is curiosity. However, there is also a fundamental element of suspense built into this type of plot, related to the reader’s expectations concerning the progress of the investigation toward a solution—that is, regarding the narrative future. But as long as the tale’s emphasis lies in the goal of the investigation, namely, the solution of the mystery, suspense of this kind is subordinate to curiosity within the overall hierarchy of narrative interest. Closure in the detective story is typically created by the successful conclusion of the investigation in uncovering all the important facts relating to the crime mystery, thus simultaneously resolving the curiosity gaps about this mystery and the suspense gaps regarding the course of the investigation.
With the basic closural mechanism operating in the detective story outlined, there still remains the question of why this closure tends to be perceived as especially strong. This perception is certainly related to the extremely “single-minded,” goal-oriented nature of the detective plot. The interest of the reader is very much focused on a single issue (the crime mystery) from beginning to end—so that when this issue is finally resolved by the solution of the mystery, the reader’s sense of relief is especially powerful. If we turn to examine the nature of the main gaps in the detective plot, which are typically related to the criminal’s identity and the whole complex of the crime’s circ*mstances, we can note two complementary aspects in which they contribute to the closural force of the detective plot’s end. The information required to fill in these gaps is usually of a simple and determinate nature. Thus, on the one hand, because of this high degree of specificity, general frames of knowledge (whether relating to literature itself or extratextual reality) cannot be of much help to the reader’s gap-filling activity. The range of uncertainty regarding these gaps remains therefore relatively wide and correspondingly intensifies the craving for additional information. On the other hand, because of the relatively simple and direct nature of the gaps, they can be straightforwardly and unambiguously closed by some determinate pieces of information once the appropriate moment arrives—however convoluted the route leading to this resolution.
Yet another group of factors that contributes to the strength of closure in the detective story is related to the special density of the retrospective patterning that typically occurs at its end. Here the solution of the mystery is not spread, as a rule, over lengthy and/or noncontinuous sections of the text but is rather packed into a relatively short section at the very end (in terms of proportion, this effect is naturally even stronger in novels than in short stories). The concentrated filling in of the gaps is usually performed (or motivated) by a detailed—and triumphant—explanation given by the detective. The solution is also supposed to be highly surprising, as a result of the reader being misled until the last stages of the story with regard to some of the most basic facts relating to the crime mystery. The strong closural effect of such a surprise twist stems from the intensity of the retrospective patterning that it forces upon us: We suddenly discover that some crucial hypotheses considered as certain—or unconsciously assumed—are in fact mistaken. But in this context, there again seems to be a complementary factor operating. Surprising as the solution to the mystery is, according to “fair-play” conventions it is also supposed to be prepared in advance by various clues spread throughout the text. Such clues are seemingly meant to help the reader solve the mystery. But actually, when one carefully examines the practice of authors in the genre, it turns out that the stricter these authors are about planting clues in their narratives the more inventive they tend to become in developing methods of misdirection aimed at preventing the reader from using the clues to solve the mystery (so as not to spoil the final effect of surprise). Therefore, a case can be made that the main function of the clues is retrospective and that their real rhetorical purpose is to lend after-the-fact credibility or probability to the solution. The clues thus enhance the closural force of the ending by providing, as they are finally unveiled, connections not made by the reader before (often while exposing connections that have been made as erroneous) and so strengthen the coherence of the plot sequence and intensify retrospective patterning as well.
The detective genre is generally considered as popular fiction, and in the current critical climate this seems natural for a genre whose typical plot structure is characterized by such strong closure. There appears to be a general tendency of modern(ist) taste to privilege openness and downgrade strong closure, the latter—and, specifically, the closed ending—being usually considered as inferior in terms of value and interest; this is compatible with the accepted notion that the open ending, and the quality of openness in general, is typical of modern literature and constitutes one of its trademarks. This zeitgeist is already clear in Joseph Conrad’s praise of Henry James at the beginning of the 20th century, for the innovative “lack of finality” typical of his works, as opposed to “the usual methods of solution by rewards and punishments, by crowned love, by fortune, by a broken leg or a sudden death” that satisfy the general reading public’s desire for conventional finality. Quite representative as well is Conrad’s justification of this new aesthetics in terms of being true to life: “[James’s] books end as an episode in life ends. You remain with the sense of the life still going on.”59 Following the authors practicing such a poetics, openness has become heavily valorized in modern criticism as well. This is true about general trends in interpretation and evaluation, as well as the more specific context of endings, with older works often being “reopened” as the attempt to demonstrate open-endedness in spite of apparent closure has become a common critical exercise.60
Open-endedness in modern narrative is a wide-ranging and complex issue, and as has already been noted, open-endedness should be considered in general not in terms of a simple dichotomy with strong closure but rather within the framework of a finely gradated and multidimensional continuum. In conclusion, here are two short observations about this modern tendency that connect with some previous points made in this article. First, endings that tend toward the open end of the continuum parallel “abrupt” beginnings, at least to a certain extent, in the historical context of their rise to prominence (the rise of modernism in the late 19th and early 20th centuries), as well as the aesthetic or poetic trend to which they testify (very generally, a desire not to accentuate the boundaries of the work of art).61 Second, it is important to note that even with clearly open endings, the strategic importance of the ending in determining the meaning and overall effect of the work often still stands. Take, for example, well-known cases of provocative open-endedness in postmodernist fiction such as Thomas Pynchon’s The Crying of Lot 49 (1966) or John Fowles’s The French Lieutenant’s Woman (1969): The refusal to resolve the main mystery in the former (with the ending reached, it would seem, just as the mystery is about to be resolved), the alternative resolutions we get in the latter, with its double ending, acquire their particular force precisely by being the last word of these novels’ authors.
Discussion of the Literature
As a point of departure for the study of beginnings in contemporary criticism one might mention Edward Said’s book on this subject, which was of considerable importance in generating interest in the topic.62 To a large extent, it is a Foucauldian critique of the concept of origins (as divine and mythical), which is opposed to that of beginnings (as worldly and humanly produced); but overall, this book is actually not so relevant to the study and understanding of literary openings. French criticism has paid quite a lot of attention to textual beginnings, particularly since the emergence of the journal Poétique, and a lot of it is condensed and systematized in Del Lungo’s book-length study; the other most ambitious theoretical study of beginnings is Leander’s volume.63 Of studies that focus on readings of specific works, two in particular are worth noting: Nuttall’s, on a series of texts from the Homeric and Virgilian epics to 19th-century novels by Dickens, its basic theme being the tensions and interplay between the sense of opening in medias res and that of a “deep,” natural beginning; and Romagnolo’s on 20th-century feminist fiction, within a theoretical framework that distinguishes between conceptual, discursive, chronological, and causal beginnings.64 The collection of articles edited by Richardson is naturally much less focused than all of the abovementioned studies, but it provides a valuable wide range of approaches and discussions of both theoretical issues and specific texts.65 Backus and Bossong offer important distinctions between types of openings based on linguistic analyses of large textual corpora.66
Two of the most important studies in the field do not declare themselves by their title to be about beginnings or openings. Sternberg’s book on the exposition studies it as the beginning of the mimetic narrative sequence, and in this capacity as a focal point of sequential dynamics involving various kinds of narrative beginnings on several levels.67 Most directly in the context of the present topic, the book’s seventh chapter focuses entirely on openings—the use of preliminary and concentrated exposition in the works of Trollope and Balzac.68 And Genette’s creation of the paratext as a critical category was instrumental to a richer and more nuanced understanding of discursive beginnings.69 It should also be noted that many studies use openings to vividly illustrate various points, without necessarily being interested in systematically exploring beginnings as such. One illustrative branch deserving a mention in this context is that of studies dealing with the construction of fictional worlds: for example, Benjamin Harshav’s presentation of his theory of the literary text’s structure, Umberto Eco on the distinction between fictional and nonfictional texts, and David Herman on the construction of a storyworld by the reader according to textual cues.70 Finally, for an extended further reading the best bibliographical sources are probably Del Lungo and Leander’s books: The former includes an annotated bibliography and the latter a chapter devoted to a historical survey of criticism on the topic.71
Regarding endings, from a certain viewpoint the amount of available material about them may be said to be enormous, since due to their strategic importance there are hardly any detailed discussions of literary texts that do not address, at least to a certain extent, issues pertaining to the ending—and/or closure—of the work. However, if we limit ourselves to studies that treat the topic from a systematic, theoretical viepoint, Kermode and Smith’s books of the late 1960s may be considered important points of departure for contemporary criticism.72 It should be noted, though, that the title of Kermode’s study is somewhat of a misnomer, since a large part of it deals with the more general topic of the basic human need to create narrative constructs in order to cope with reality; within this framework, the ending or conclusion constitutes only one element—albeit of major importance—among a whole array of patterning devices activated by a narrative. Smith’s study, although more than half a century old, is arguably still the most important in the field in view of its contribution to the concrete understanding of many ways in which closure is produced—or not—by the endings of literary texts.
Following Smith’s study, which focuses on poetry, attention in this line of investigation shifted to narrative, and the most detailed later studies of closural devices are those by Korte and Krings.73 These tend, however, to become enumerations of such devices (though they also deal with the historical development of their use), losing sight of the holistic aspect of Smith’s approach, which examines closure with an eye to the structural dynamics of the text as a whole. A study of a specific genre that preserves this orientation is David Richter’s Fable’s End, on idea- or thesis-driven narratives.74 Richter explores how the didactic element of the texts he is analyzing influences their narrative structure, focusing particularly on the problems of closure these structures give rise to (closure not just on the story level but also in substantiating the narrative’s thesis), and the main strategies employed to deal with such problems. Another good genre-specific discussion is the chapter on the dynamics of closure in Janet Altman’s study of epistolary fiction; a special claim to notice is that because of the nature of episotolarity, a particularly great emphasis lies here on the dynamics of the communicative act in addition to (or even instead of) that of the narrated action.75 Richter’s study of thesis-driven texts also points to another line of investigation, with an emphasis on the ideological aspect of ending and/or closure. Two studies that deserve a particular mention in this context are Rachel Blau Duplessis’s Writing Beyond the Ending, which examines the endings of several 20th-century texts that transgress the conventional fates of female protagonists in marriage/suicide/madness, thereby expressing critical dissent from 19th-century models and attitudes; and Russell Reising’s Loose Ends, which reflects on how concluding moments in various representative American texts accentuate ideological fissures that these texts seem to suppress.76 The articles by Carroll and Segal present general models of narrative closure (Segal’s also includes a detailed illustration from detective fiction), outlining holistic approaches that recognize the text’s overall dynamics and defining closure in terms of audience perception.77
Of studies that discuss issues related to ending and closure with a focus on specific corpora, it is worth further noting Miller on the 19th century (“traditional”) novel, Torgovnick on various 19th- and 20th-century novels by authors from Dickens to William Faulkner, John Gerlach on the American short story, Armine Kotin Mortimer on the French novel, and finally several books dealing with pre-novelistic corpora: the collection edited by Deborah Roberts, Francis Dunn, and Don Fowler on ancient Greek and Roman literature, Troy Troftgruben on the problematic ending of the Book of Acts in the context of ancient literature, and Susan Zeelander on biblical narrative (mainly the Book of Genesis).78 Such early corpora may raise special methodological issues in the context of the study of ending and closure—for example, how to segment narrative units within the biblical text?
