The theory a writer needs for arranging times answers four practical questions: what is being rearranged, why the same events mean something different in a different order, what keeps a reader going across a change of time, and what that change costs her. Each strand below opens with the question and then gives the source.
What is being rearranged: two orders and the anachrony
The question. When a writer moves a chapter from the end of a book to its beginning, what has she changed, and what has she left alone?
The Russian Formalists gave the first answer. Viktor Shklovsky and Boris Tomashevsky, writing in the 1920s, separated the fabula, the events in their causal and chronological sequence, from the syuzhet, the same events as the work arranges them (Tomashevsky’s “Thematics”, 1925, translated by Lee T. Lemon and Marion J. Reis, 1965). Moving the chapter changes the syuzhet and leaves the fabula alone; the writer has changed nothing that happened and everything about how it is met. This chapter calls the two story order and text order.
Gérard Genette’s Narrative Discourse (1972; translated by Jane E. Lewin, 1980) made the comparison precise. He called any discrepancy between the orders an anachrony, an analepsis when the text goes back to an earlier event and a prolepsis when it goes forward to a later one, and measured each by its reach, how far from the present it goes, and its extent, how much story time it covers. He also named a limit case, achrony, an event whose position in story time cannot be determined from the text, which is where a mosaic of fragments ordered by association ends up. A writer charting a draft uses the same notation without the vocabulary: number each unit by its place in the story, then list the numbers in the order the text gives them. That list is the order chart of The Mechanism.
Why order changes meaning: primacy and recency
The question. Why does a reader who meets the murder first read the marriage differently from one who meets the marriage first?
Menakhem Perry’s “Literary Dynamics: How the Order of a Text Creates Its Meanings” (1979) argued that a reader builds a frame from what comes first and reads what follows through it, the primacy effect, until the text forces a revision, after which the most recent frame governs, the recency effect. Meir Sternberg had made a related argument about exposition in Expositional Modes and Temporal Ordering in Fiction (1978): where a writer places the facts a reader needs decides which of them she treats as the norm and which as departures. The Prologue (Chapter 10) uses the primacy effect for prologues. For a non-linear book the consequence is that every arrangement installs some frame first, and the writer should choose which. Chronicle of a Death Foretold (1981; translated by Gregory Rabassa, 1982) names the killing in its first sentence and then goes back to the hours before it; the reader reads every delay and every warning not given as a step towards the death, which is the frame Gabriel García Márquez wants.
What keeps a reader going: suspense, curiosity and surprise
The question. A switch takes the reader away from a strand at the moment she cares about it. Why does she keep reading?
William Brewer and Edward Lichtenstein’s structural-affect theory (“Stories Are to Entertain”, 1982) explained the three pleasures of plot by the relation between story order and text order. Suspense arises when the text gives an initiating event and withholds its outcome; curiosity when it withholds an earlier event and lets the reader know it is missing; surprise when it withholds an earlier event without the reader knowing, and discloses it late. All three are made by ordering, which is why an arrangement of times is a machine for producing them. A switch at the right moment converts a strand’s open question into suspense held across the other strand; a strand set later than another produces curiosity about how the earlier one gets there. In their experiments readers judged texts more story-like when the order produced one of these effects and the outcome was finally resolved, which The Reveal (Chapter 15) relies on as well. The theory also predicts the commonest complaint about braided books: a switch made when the reader holds no question about the strand she is leaving produces no suspense, and the strand she enters must earn her interest from nothing.
What a change of time costs: the evidence
The question. How much does a jump in time cost a reader, and does a non-chronological book cost more?
Rolf Zwaan’s event-indexing model of reading (Zwaan, Langston and Graesser, 1995) holds that readers track a situation along several dimensions, time, place, cause, the characters’ goals and the protagonist, and update their model when any of them changes. Zwaan’s “Processing Narrative Time Shifts” (1996) measured the time dimension. Sentences that opened with An hour later or A day later were read more slowly than ones that opened with A moment later, and information from before the larger shift was recognised more slowly afterwards: a shift in story time, not the number of words, made earlier material less accessible. Keisuke Ohtsuka and William Brewer (1992) compared texts told in chronological order with texts that told events out of order, and reported that readers understood the order of events less well in the second. The evidence is from short texts read in laboratories, not novels, and the chapter uses it for one modest conclusion: every switch between times costs the reader an update, and a book that switches often must pay for each with a marker she can read at once and a reason to make the effort.
The convention
Chronological order is itself a convention, and not the oldest. Horace advised epic poets to begin in medias res, and the Odyssey tells Odysseus’s wanderings in his own retrospective account in the middle of the poem. The realist novel’s preference for following events in sequence, with cutaways where needed, dominated the nineteenth century and was broken deliberately by modernists such as William Faulkner and Virginia Woolf. Whether multi-timeline novels are commoner now in commercial fiction than they were, as is often said, this chapter has not measured, and it treats the claim as an impression. The tools below assume a reader who reads a book once, in the order printed, which is the reader every arrangement is designed for.
See also: the full treatment, section 2, with Genette’s notation worked on an example, the formalists at greater length, and Brian Richardson’s accounts of temporalities that no story order can reconcile.