Narrative is an art of time; it uses the ordered sequence of linguistic signs as a natural sign, an icon, of the sequence happenings in the action. We shall take the chronological order of representation as the standard, unmarked one. When tell a story we tend to move from beginning to end, using representational time as an icon of the represented time:
Action time a b c d e f g
Story time A B C D E F G
In principle, therefore, stories move forward, in an iconic way which signals the passage of time. But this is only the general rule. Stories may suddenly jump back, against the direction of temporal progression. They may also jump forward, interrupting their normal pace, or move in a variety of speeds in one direction or another, compressing or expanding the narrated time. The Bible, for instance, moves from Creation to Apocalypse mirroring the temporal sequence of the events it tells. Note, however, that the Gospels retell the same events up to four times‹ which is just one instance of a break in one-to-one sequence. Chronological order is our unmarked standard, but it is still a conventional one. Nearly all stories will modify the standard order in some way or other, though few are as systematic as Martin Amis's Time's Arrow, which moves steadily backwards from the moment of the hero's death to his birth and beyond. Notice, however, that even in this case narrative time is employed as an icon, a pictural sign, of narrated time_ albeit a partially refunctionalised one.
So, there are two basic story orders: the simple, unmarked order of chronological succession of events, and the complex order that includes some kind of temporal distortion. This was also the assumption of medieval rhetoric, which distinguished between the natural order of telling, ordo naturalis, and the more artificial, or artful, ordo artificialis. The story can distort the order of the events in various ways. Any temporal distortion which modifies the chronological sequence of events is an anachrony.
Anachronies have always been common in literature. In fact, Aristotle seems to have been the first one to make a remark on this phenomenon, when he compares the temporal structure of the tragedy and the epic:
In tragedy we cannot imitate several lines of actions carried on at one and the same time; we must confine ourselves to the action on the stage and the part taken by the players. But in epic poetry, owing to the narrative form, many events simultaneously transacted can be presented (Poetics 63, XXIV.4)
Of course this feature is not a necessity, but a convention of the Greek stage. However, it does suggest that linguistic narrative easily yields to temporal distortions, and that the time scheme of a novel will usually be more complex than that of a play (or a film). The study of anachronies was undertaken by German and Russian Formalists (Käte Friedemann, Boris Tomashevski). The most complete system is expounded by Gérard Genette, and we shall frequently refer to his system.
Anachronies are determined with respect to a reference point: the ongoing story-time (the "main story"), or the immediate events narrated before the anachrony is introduced. Note that this reference point is not the moment of narration, the moment in which the narrator speaks. Similarly, we know that the anachrony is over when the interrupted sequence is resumed, and we return to the normal unfolding of the action. Genette calls that unfolding of events the "first narrative". His definition is "the temporal level of narrative with respect to which anachrony is defined as such" (ND 48). In standard narrative, the beginning of the story usually marks a significant moment in the action; although later on we may go backwards for expositional information, the main chain of events is usually tightly knit from the point marked by the beginning of the story. It goes without saying that the relative coherence of the first narrative will reinforce the subordinate character of the anachronies. If no coherent first narrative is formed (as in Molly Bloom's monologue) there results a temporal constellation in which every element is defined and defines the others in equal measure. In experimental narrative it may be impossible to determine a "main story" sinche the causal and chronological links may blur any clear-cut outlines, and there is no stable reference point from which anachronies are consistently determined: rather, one anachrony may be grounded on another and the process can be repeated indefinitely.
We have defined an anachrony as a temporal distortion between the time pattern of the action and the time pattern of the story. There are several possible criteria of classification: order, reach, extent, diegesis, level, motivation...
Considered in terms of pure ordering of events, there are two kinds of anachronies: an anachronical event may belong either to the past or to the future with respect to the events which form its immediate context. We call the first type analepsis or flashback; the second type is prolepsis or flashforward.
A flashback is a break in the sequence which introduces events previous to those being told in the narrative sequence which has just been interrupted:
Action time: a b c d e f g
Story time: C D E A B F G
A flashback may be used to introduce missing information from the past. But of course the events may be told in the unmarked order and then repeated at a later stage:
Story time: A B C D E A F G
The flashback will simply have a different mission, one of underscoring the significance of those events, showing them in a new light or maybe just reminding them to the reader.
A flashforward is the mirror image of a flashback. It will introduce events which are "future" with respect to the reference point of the immediately surrounding sequence.
Action time: a b c d e f g
Story time: A B F C D E G
These events, however, may already belong to the past for the narrator. The whole of the action schematised above might be over before the narrator begins to tell the story, so the flashforward in F is just as past as the events in A or B.
Prolepses are less frequent than analepses, although they are perfectly coherent while we remain in retrospective narrative. But when they are present they also contribute to the structure of expectation, curiosity and suspense, to the activity of gap-filling and construction of coherence which is the task of the reader of narrative. Sometimes an otherwise straightforward narrative may include prolepses which accentuate the feeling of curiosity: how shall we reach the stage adumbrated by the prolepsis? A novel with a complex temporality such as Salman Rushdie's Midnight's Children makes a constant use of this kind of curiosity-goading prolepsis.
It is important to realize that these temporal distortions have to be apprehended at some point as we move through the story. We cannot define the temporality of the story as a simple formal scheme: time must enter the description. From the moment the reader constructs a coherent series of events he has a temporal orientation and a "now" moment; any anachrony will be perceived to be a flashback or a flashforward with respect to that moving present. It is important to understand the nature of this definition: anachronies are not measured with respect to the time of the enunciation (as, for instance, verbal tenses) but with respect to a narrative reference point created by the ordered unfolding of events up to that moment.
Reach and extension
Anachronies may vary in reach and extension. The reach of an anachrony is the temporal distance between the reference point and the anachrony itself. The anachrony may go beyond the limits of the main story: for instance, a story which takes place during a few days may contain a flashback which tells events several months or years back in the past. The extension is the amount of time presented in the anachrony: from a single event to whole sequences or to the greatest part of a narrative.
Both prolepses and analepses can be external or internal (with respect to the beginning and end points of the main story) and have two relevant dimensions: reach and extent. They may also be homodiegetic or heterodiegetic, that is, dealing or not dealing with an action line which is narrated earlier or later in the main story. Heterodiegetic anachronies are still somehow related to the story since they refer to events set in the same narrated world (otherwise they would not be mere anachronies), but those events are not causally tied up to the main story. Intermediary cases are of course possible: for instance, narratorial flashbacks which inform us of the previous life or habits of a character when it first appears onstage.
Internal homodiegetic analepses are used to recapture previous action material. They may add something new or just repeat previous information. Repeating analepses, or recalls, tie the narrative to its own past and, if they do not add to the narrative information, can be an important principle of stylistic construction. Completive analepses, or returns, "fill in, after the event, an earlier gap in the narrative." This play of creation and filling in of gaps contributes to create a specific kind of narrative interest.
Narrative is a transformation not only of the action material, but also of the reader's impressions; both returns and recalls help effect this transformation. It is important to note that the difference between recall and return is not a clear-cut one. New aspects of a phenomenon may appear in a later recall; the same event may be modulated in a different way through the attitudes of the narrator or the characters. Here as elsewhere, the concepts we introduced should be used as measuring rods rather than pigeon-holes.
We can divide objective anachronies (or story anachronies) from subjective anachronies (or action anachronies). Subjective anachronies are a part of the action, and are as independent of the narrating instance as any other event: they are a part of the characters' subjective psychological activity. Objective anachronies are the product of the extradiegetic narrator's deliberate manipulation of story time.
Subjective anachronies are a typical case of realistic motivation: the information in the story is ordered with a view to the effect on the reader, but the surface justification for the introduction of previous materials is the remembrance in a character's thoughs or narration. A subjective anachrony is not as obvious a distortion of narrative order as an objective anachrony, since in one sense it never disrupts the order of the events: it merely presents some of these events not directly but mediated by signs which repeat them further on in the story, a tale, a remembrance. The surface order of events is left unaltered.
Subjective anachronies are for the most part analepses, since subjective prolepses are obviously more difficult to motivate realistically. Objective prolepses are one of the most intrusive ways in which a narrator may control the reader's expectations, to deflate his expectations about the plot. In practice they are used very rarely even in the most intrusive narratives: narrative sequence cannot stand a continuous disruption of anticipations. Prolepses of this kind seem to defy one basic principle of narrative: why go on telling if you can tell the ending straight away? They also lay bare a related unstated narratorial privilege: the narrator can feign his knowledge of the story unfolds gradually, that he does not withdraw information gratuitously ‹ as long as he does not choose to offer information gratuitously, in defiance of the gradual unfolding of the tale. Some realistically-minded critics condemn this procedure outright, as it seems to determine the fate of characters in advance and to assert an excessive distance between the author's privilege and the reader's passive stance. Of course, this resource, as any other narrative figure, can be used in a wide variety of ways ‹ incidentally or crucially, seriously or parodically. Here we only outline some of the most predictable and frequent effects it creates.
Just as a straightforward narrative may contain embedded anachronies (subjective anachronies), these anachronic sections may in turn embed third-level anachronies, either proleptic or analeptic with respect to both the moment of their insertion and the moment of insertion of the first subjective anachrony. This phenomenon is of course related to the several types of embedding we discuss at lenght elsewhere: in its purely temporal aspect we may refer to it as temporal embedding. This multiplication of references in the direction of both past and future may complicate the temporal structure of a tale considerably; they may even give rise to achronic sequences.
We frequently find two well-defined points at which the anachrony is articulated to the main story: the beginning and the end. An analysis of the anachrony should include an analysis of the way this articulation takes place. Which, is for instance, the motivation for the anachrony? Is it the introduction of some new character? Is it necessary for the unveiling of some secret? Like other story devices, anachronies may be motivated at action level, for instance using a character's remembrance of the past (this has a bearing on the level of focalization), or the finding and reading of an intradiegetic report (involving a shift of narrative level). They can also be motivated at the level of the fictional narration: for instance, a first-person narrator may withhold information till the end in order to procure mystery and suspense for his narrative. They may also be pure story anachronies, without a motivation coming either from the narrators or the fictional world. This is the case, for instance, in Faulkner's The Sound and the Fury, where the story is divided in four nonsequential sections. The order of those sections is the author's responsibility, since each of the sections has a different narrator who is responsible only for the anachronies within his section, not for the articulation between sections.
The ending of the anachrony is also relevant. It may be well defined, or it may shade into a different narrative mode. An analepsis may even catch up with the main narrative without a break in continuity, by telling us an alternative sequence of events which leads to the reference moment‹ these are complete analepses in which reach and extent have the same duration. A significant analysis of the anachronies will relate them, for instance, to the development of the plot. Are the anachronies events in their own right, that is, are they a part of the action, and not merely a part of the narrative discourse? A flashback may be a confidence between the narrator and the reader, without any of the characters taking responsibility for it or knowing of it. Or it may be a speech event in the story, with a character telling another about the past (e. g. the old servant's narrative in Wuthering Heights). It will be useful, therefore, to watch for the precise narrative articulation of the anachrony, to define its nature from the point of view of other narrative categories. Narrative level, for instance. An anachrony may coincide with a change in narrative level (e. g. a character speaking) or it may be a temporal shift without a shift in voice (e. g. a character telling a story suddenly explains a previous event to clarify some point in the story he or she is telling). The same goes for other narrative categories (aspect, mood). Just saying that a piece of discourse is an anachrony does not give us a precise idea of its temporal structure, because temporality is highly complex, and is not a mere matter of sequence. Subjective time, or durée, for instance, related to the point of view of one character, may not coincide with the experience of time as presented by the narrator, and a description of the contrast in perspective is necessary to fully characterize the temporal contrast.
The status of an anachrony should also be taken into account. A flashforward may be a real prediction, or it may just be a piece of wishful thinking on the part of a character: that is, its status can be either factual or hypothetical. Most elements in the action have a simple temporality, being merely signified events. But some of the action elements (objects or events) are signs, and as such may have a double temporality: the temporality of the signifier and the temporality of the signified. Therefore, the temporal status of these elements will have to be described at two levels of signification: the standard semiotic level of the story and the signified semiotic level of their referent. For instance, an epic narrative may suddenly give way to a description of past events which are depicted in a present work of art portrayed by the narrator (e.g. Aeneas looking at paintings of the destruction of Troy in Dido's castle). Or a character may reminisce through a story: the telling of the story, the story-as-sign, is located in the present; but the events depicted in the story take us to the past. There is no anachrony in one sense, since the present goes unfolding itself. But there is an anachrony in another sense, since we learn about the past or the future.
That is, "real" anachronies can be introduced by the narrator, but they are not the only possible ones. There are also anachronies present in action elements (speeches, stories, works of art, memories) which are capable of signifying an autonomous temporality. A distinction between objective versus subjective anachronies (Genette 1983: 47) is useful, but it remains too general. Moreover, the term "subjective anachronies" refers only to anachronies introduced by speech or psychical processes. Maybe it is better to speak of action anachronies as opposed to story anachronies. In the last chapter of Ulysses, which contains Molly Bloom's interior monologue, there are no story anachronies: the rhythm and sequence of the mode of presentation, a sequence of thoughts, are uninterrupted. But there is a complex anachronical structure in the contents of those thoughts. This ability anachronies have to contain other anachronies inside them can greatly complicate the temporal articulations of the story. A prolepsis can contain a second-level analepsis which contains a third-level prolepis, and so on.
Anachronies must also be studied as they interact with questions of narrative voice. First-person narrative could be defined as intrinsically analeptic: the narrator reaches back from the present moment of enunciation to the past moment in which the story unfolds. But each narrative may emphasize or de-emphasize this fact in a specific way. Any first-person narrative can ideally end as a complete analepsis, that is, with the story reaching the temporal point of narration, as a snake swallowing its own tail. This use of temporal structures is exploited in a novel such as Molloy for metafictional purposes. But it is disregarded in most first-person narratives.
Reading anachronies: suspense, surprise and curiosity
Needless to say, anachronies are a powerful shaper of the reading experience, the way the reader constructs the narrative and the specific kind of emotions produced by the narrative. Pushed to the extreme, chronological order keeps the reader completely informed of the progress of the action: there is no need of coming back to retake some unexplained event; everything has been told and therefore the attention of the reader is riveted on the future, not on the past. The peculiar emotion produced by this kind of straightforward narrative is suspense: the reader wonders what will happen, and the whole of his interpretive attention is projected to the future. The model for this kind of narrative is perhaps the adventure story: war tales, westerns, science-fiction, children's tales...
If we push the second variety of narrative ordering to its logical conclusion, we find that here the logic is double: as in straightforward narrative, we wonder what will happen next, but, since important facts are being concealed from us for the moment, we also wonder what has happened. That is, curiosity is the reader's main passion here, or curiosity combined with suspense. We wonder about the nature of the past in order to explain the present, but we also wonder about the way in which the past will be revealed, the revelation of its full hold on the present. The prototype for this kind of story is the detective story, which unfolds simultaneously toward the origin and toward the conclusion of the action. This second kind of story can't be content with a simple, one-way progression into the future. It needs to come back on itself, and finish what was left unfinished, tell us the mystery which has been hidden all through the story. A temporal distortion is needed, the most basic one, a return to the past which will enable us to understand the present.
Not all informational gaps in narrative structures are used to create mystery. Curiosity only arises if the reader knows about the existence of such gaps. But we have to distinguish these curiosity gaps from surprise gaps,  whose existence is only revealed at a later point in the development of the story, and are used to give an unexpected twist to its development.
Vertical Integration: Exposition and Motivation
Tomashevski has introduced a series of terms to describe the "horizontal" sections of a action: initial situation, exciting force, peripeties, climax, ending. To describe reading effects we need a further term, the exposition, which is "a narrative introduction to the initial situation", or "the presentation of circumstances determining the initial cast of characters and their interrelationships." The exposition is not the beginning of the story. It is distinguished in a footnote from the parts of the siuzhet, the beginning and the end as they are actually found in the work:
From the point of view of the arrangement of the narrative material, the part beginning the narrative is called the prologue. The close is called the epilogue. (1965: 72)
The exposition cannot be defined with respect to the siuzhet alone. But it is not simply a part of the action either. Otherwise, we would just call it "initial situation." The exposition is, therefore, the representation in the story of the initial situation of the action It is a concept having to do not only with cause and effect (since it usually deals with static relationships or with the causes of a particular situation) but also with the informative structure of the work: the exposition is there to serve the artistic interest and control the reader's knowledge, while allowing him to acquire the necessary causal and temporal orientation in the narrated world. Its definition must therefore be intrinsically relational, half-way between action and story. In this way, Tomashevski avoids the inadequacies which Freytag's definition of exposition as the initial part of the work presents for an extended narratology.
However, we should not forget that the beginning of a text is always expositional to some extent. It may not acquaint us with the central facts of the story, but it always installs the reader in the narrated world, making him acquainted not just with characters, settings, etc. but also with narrative style and technique. Analysis should take into account the relationship between exposition proper and the textual opening.
Tomashevski divides exposition into immediate, delayed, and transposed. The latter involves a time shift for one whole section of the narrative, while delayed exposition is gradual and correlative with an ex abrupto beginning. The concept of delayed exposition leads Tomashevski to deal with point of view and the use of focalizer characters: "Usually the author withholds information about the circumstances involving a group of major characters, telling the reader only what one or the other of the characters knows" (1965: 73). Let us note here that the important thing is not what the "author" (in fact, the narrator) knows, but instead what he says to the reader.
We have mentioned two prototypical or ideal kinds of story, depending on whether they follow the logic of succession or the logic of retrieval and completion of information; the adventure story and the mystery story. Since we define these two kinds of story with respect to the kind of expectation they arouse in the reader, it is obvious that a formal description of a story has to take into account the temporal development of the story: a story is not only what it 'really' is, but also what the reader thinks it is when it is being read. A suspense story might reveal itself in the end as a mystery story, and it is this succession of expectations in the reader which provides an adequate account of its form.
A mystery story necessitates that the reader ignore part of the action. This can be achieved in various ways. The story can begin in medias res, and the delayed exposition appears gradually later on. These stories are born with a mystery in them. However, the mystery may develop during the unfolding of the story. It consists then in a control of the information available to the reader. A mystery can be described as a gap in our knowledge of the story. A mystery story is therefore a system of creation and resolution of informational gaps. Gaps can be divided into permanent (unsolved mysteries) or provisional. The nature of gaps can usually be determined objectively, but the analyst must take into account the impression of the reader: a gap which normally belongs to the class of provisional gaps can sometimes be left open forever (cf. John Fowles's story "The Enigma", in The Ebony Tower). So, we can establish an opposition between provisional and permanent gaps. On the other hand, we can take into account the reader's perspective in order to distinguish between curiosity gaps, those which are recognized immediately, and surprise gaps, informational restrictions which only reveal themselves in their full extent from a later perspective. Among surprise gaps we may mention paralipsis, a piece of information which suddenly reveals itself to have been skipped while it should have been available under the existing mode of presentation.
Actually, the two kinds of narrative are extremes, and most narratives combine both kinds of interest. And that is because, from its very definition, narrative has two main movements; it looks both to the immediate future, following the logic of succession, and towards some point in the past, following the logic of repetition‹any narrative is in a sense a repetition of the events it tells.
Ordering of different actions
The ordering of different actions or action-lines should be analysed taking into account the temporal relationship between the inner development of these story lines and their articulation. For instance, an alternance of two story lines may use as breaking points a moment of crisis and high expectancy, or else a lull in the action. The intervening action line may be used iconically to represent elided action-time in the alternating action, or else it may lead to that action being retaken at the very moment in which it was left. If the alternating action is introduced at a secondary narrative level, the temporal consequences of this motivation may also be relevant. The new action may be a narrative read by the hero or told to him (as in "El curioso impertinente" in Don Quijote or Nelly Dean's tales in Wuthering Heights, or it can assume the form of a picture or a dream.
There are three possible relationships of order between different action or action-lines: sequential enchaining, embedding (whether enunciative or otherwise), and alternance. If the episodes of an action are quite independent as far as causality is concerned, if few or only one character is the common element, if they follow each other sequentially, we may find hybrid forms between the novel and the collection of short stories. The connecting link may be a community (Dubliners, La colmena, Winesburg, Ohio, Go down, Moses), the main character (Beckett's More Pricks than Kicks) or a set of themes and images (The History of the World in 10 1/2 Chapters). Alternating or embedded actions provide a tighter, more encompassing structure for the episodes.
 Cf. Lessing 1985: 148ff.
 See Fowler 1977: 6-7.
 See Genette 1972: 79.
 Genette 1980: 51.
 This is clear from Genette's examples, e.g. 1980: 58.
 Cf. Genette 1972: 89; Bal 1977: 117ff; 1985: 64.
 See Trollope, Barchester Towers 129-30; Lawrence Durrell, Justine 66-67.
 See Pouillon 1970: 73. On the other hand, Booth sees it as a legitimate way of achieving ironic distance, of "controlling the reader's expectations, insuring that he will not travel burdened with the false hopes and fears held by the characters" (1961: 173).
 Genette 1972: 118.
 Sternberg 1978: 244-45.
 See Meir Sternberg's critique of Freytag in Expositional Modes and Temporal Ordering in Fiction (ch. 1).
 Sternberg 1978: 238 ff.
 With Sternberg 1978: 244f.
 Cf. Genette 1980: 52.
 Cf. Shklovski 1965b: 196; Propp 1971: 108ff.; Todorov 1966: 140.