The deficit claim in the problem file, which the Reference chapter calls the interest deficit (the attention a cutaway spends before it has delivered anything), is a metaphor, and a writer revises better when she knows what it stands for. Five strands of theory answer five practical questions.
Why does going back cost more than going forward?
Research on reading comprehension gives the cost a mechanism. Rolf Zwaan and Gabriel Radvansky’s review of the field (Psychological Bulletin, 1998) holds that readers do not store sentences but build a situation model, a mental representation of the state of affairs the text describes, tracked along five dimensions: protagonist, space, time, causality and intention. A shift on any dimension forces an update, which shows up as slower reading. Catherine Emmott’s Narrative Comprehension (1997) adds the reader’s contextual frame, the running record of who is present, where and in what configuration, and distinguishes switching to a new frame from recalling an old one. A flashback needs both and something more: the present frame is not abandoned but held in suspension while a second is built, and then recalled at the return. Keiko Ohtsuka and William Brewer (Discourse Processes, 1992) found that readers reconstruct the underlying chronology of a story told out of order, and answer questions about its order less accurately than when the same events are told in sequence.
That a backward shift costs more than a forward one of the same size is plausible on this account, but the evidence for a general asymmetry is suggestive, not settled. The findings describe comprehension, not value: they predict that Absalom, Absalom! is hard to follow, and say nothing about whether it should have been written otherwise.
What is the reader leaving, and how far?
Genette’s first narrative answers the first half. Since anachrony is measured against it, a narrative needs a stable first narrative before a departure can be felt as one; a long flashback in a book’s second chapter is read not as a departure but as the book not yet knowing what it is.
The second half is answered by two measures. Reach (Genette’s portée) is how far back the cutaway goes from the present of the first narrative; extent (amplitude) is how much story time the cutaway itself covers. They are independent, and the table sets out the four combinations.
| Short extent | Long extent | |
|---|---|---|
| Short reach | Yesterday’s argument, dramatised: cheap and underused | The last six months of a marriage, summarised: cheap, the natural home of summary |
| Long reach | One afternoon thirty years ago: the most expensive cell | A childhood or a war: in effect a second book, needing an architecture |
The long reach with the short extent is expensive because the reader must build a whole vanished context, other ages, another place, a superseded set of relations, for one scene’s use. That ratio of construction to occupancy, not anything about childhood, is what lies behind the workshop’s suspicion of the childhood flashback. Reach can also change while a cutaway runs: a long analepsis that moves forward closes the distance to the present, as Nelly Dean’s chronicle in Wuthering Heights starts about thirty years before Lockwood’s present and runs on until it meets it, and the reader feels the closing as momentum.
Genette’s further distinctions locate the trouble. An external analepsis lies wholly before the first narrative begins and cannot contradict anything the reader has been shown; an internal analepsis falls inside its span and risks repeating or contradicting it; a mixed analepsis starts outside and runs in. Within the internal case, a completing analepsis fills an ellipsis the reader knew was there, and arrives as an answer, while a repeating analepsis returns to material already narrated in order to recolour it, and is worth the reader’s time only if the second telling differs. The Good Soldier and Beloved are built almost entirely of repeating analepses and survive because each return differs from the last, which is the warning against reading the cost model as a prohibition.
Can the cost be paid in advance?
Sternberg’s Expositional Modes and Temporal Ordering in Fiction (1978) answers this. He distinguishes exposition that is preliminary (delivered up front), delayed (action first, antecedents later) and distributed (dispersed and never gathered), and three interests a narrative can raise. Suspense is uncertainty about the future, discharged when the outcome arrives; curiosity is uncertainty about the past, discharged when the gap is filled; surprise is the discovery that there was a gap at all, opened and filled at once. A cutaway that answers a question the reader is already asking is not an interruption but a payment. The household novel uses this: from the drawer onwards the reader wants to know why the looser hand in the ledger begins on 9 March, so when Dov’s chapter goes back to the evening of 9 March (The Viewpoint Shift, Chapter 13) it pays a debt the book contracted a long way earlier. Hence the rule: open the gap before you fill it. Prolepsis does the same work from the other side, since a reader told what will happen wants to know how; the death announced at the start of The Death of Ivan Ilyich turns the whole retrospect that follows into the completion of a gap.
When is the past the story?
Paul Ricoeur’s Time and Narrative (1983–85; translated by Kathleen McLaughlin and David Pellauer) supplies the counterweight: locally a cutaway interrupts, but at the scale of the book the arrangement of past against present is the book’s meaning. Trauma theory pushes further. Cathy Caruth’s Unclaimed Experience (1996) describes traumatic experience by its belatedness: an event not fully registered when it happened returns intrusively, so that it cannot be told in sequence and possesses the present rather than preceding it. Sethe’s rememory in Beloved, a past that persists as a place someone else can walk into, is its fictional form. Caruth’s model is contested, by Ruth Leys (2000) on its clinical basis, by Dominick LaCapra (2001) on its collapse of acting out into working through, and by Stef Craps (2013) on its Eurocentrism; for a writer the upshot is the same either way. The conventions of fragmentation and return exist and readers recognise them, and a book that borrows them without the subject, fragmenting material that is merely withheld, has taken on obligations it cannot meet.
The dates of the convention
The word flashback is a cinema term, and Maureen Turim’s Flashbacks in Film (1989) shows how its conventions, the dissolve, the voice-over, the return to a face, were worked out in one industry over a few decades. Earlier narrative held its past differently. When Eurycleia recognises Odysseus by his scar in Book XIX of the Odyssey, Homer breaks off for some seventy lines to tell how he got it, and Erich Auerbach’s Mimesis (1946; translated by Willard Trask, 1953) argues that nothing is suspended while he does, because Homer’s narrative has no background held in reserve. The modern cutaway, with its motivated trigger and clean return, depends on a suspended foreground, and the prohibitions this chapter teaches describe a convention about a century old, reinforced by workshop teaching. They describe a certain kind of contemporary novel, not narrative.
See also: the Reference theory section, with Mendilow, the historical forms of the past from the epic excursus to the oral chronicle, and the interest deficit restated with its four refinements.