6. Braiding and Convergence

6.1 Two clocks, one spine

A book has one spine. Two chains running concurrently in story time must nonetheless be delivered to the reader in sequence, and the sequencing is a design problem with no natural solution — because nothing in the story determines where the cuts fall.

Two operations are involved, and craft discussion routinely conflates them.

Braiding — the interleaving of two or more chains in discourse time: how often the narrative changes thread, at what points, and in what proportions.

Convergence — the relation the chains eventually bear to one another: whether, when and how the secondary chain intersects with, complicates, or resolves into the primary one. Convergence may be literal (the chains meet in event) or thematic (they never touch, and comment on one another by juxtaposition).

Braiding is a question about the reading experience; convergence is a question about the story’s architecture. A book can braid beautifully and converge not at all (§8.1), or converge decisively after a braid so sparse that the convergence is unearned (§8.5). The two must be designed together, and §6.6 is where the chapter says how.

One further distinction, borrowed from Genette, prevents a common confusion. Narrative cannot represent simultaneity; it can only alternate, and the alternation is a property of the discourse, not of the story. Two threads that a reader experiences as running “at the same time” are in fact being read one after another, and every device that produces the impression of simultaneity — the cut at a moment of suspension, the overlapping clock, the shared event seen twice — is a discourse technique with a cost. Chapter 5 (The Transition) treats the costs; this chapter treats the pattern.

6.2 Historical grounding: where braiding came from

Braided narrative is old and was not invented by the novel.

Interlace. Medieval romance, particularly the French prose cycles and their English inheritors, used a technique the scholarship calls entrelacement: the narrative follows one knight until a point of suspension, leaves him, follows another, and returns, with the abandonment signalled by an explicit formula — now the tale leaves off speaking of X and returns to Y. Malory’s Morte Darthur preserves the formulas. Two features of the medieval practice bear directly on modern craft. The transitions are announced, which removes the orientation cost entirely at the price of the illusion of continuity. And the threads are held together by a shared quest object rather than by causal linkage — the Grail integrates threads that never meet.

The Renaissance stage. Levin’s material (§3.2). The theatre’s braiding constraint was the scene, and its integrating devices were analogy and social hierarchy rather than causality.

The serial novel. The nineteenth-century monthly number is the form in which braiding became an industrial technique. A part had to contain variety, end at a point of suspension, and remind readers of threads last seen a month ago. Dickens’s surviving number plans — the working notes in which he allocated threads and chapters across each part — show the allocation being made deliberately, thread by thread, and they are the best documentary evidence available for how a braided novel was actually managed. Chapter 6 (The Chapter Break) treats the number plans in detail.

The twentieth century and after. The modernist novel largely abandoned braided plots while intensifying braided consciousnesses: Mrs Dalloway alternates minds without alternating chains. The braided plot survived in genre, in the saga, and in the historical sequence, and returned to cultural dominance through long-form television, where the A/B/C structure is a production requirement (§10.4).

6.3 What the scholarship gives the writer

Three of the theoretical sources in §3 convert directly into design instruments.

From Levin: a classification of the relation. Every pair of chains stands in a relation of analogy, contrast, direct linkage, or some combination. Naming the relation is the first design decision, and it is prior to any decision about braiding, because the relation determines how visible the crossings must be. Direct linkage can survive long dormancy, because the causal connection reasserts itself on resumption. Analogy cannot, because analogy exists only in the reader’s active memory.

From Empson: the requirement of proximity. If meaning is produced by juxtaposition, the two terms must be juxtaposed — which in a novel is a fact about page distance, not about thematic relation. This is the strongest theoretical argument for tight braiding in analogically related books, and it is not usually stated in craft literature.

From Chatman: a definition of the convergence point. The moment at which the subplot’s kernels become the main plot’s kernels (§3.4). This gives the writer something to look for: not “where do they meet?” but “where does deleting this break the main story?”

6.4 A taxonomy of braiding and convergence structures

Braiding and convergence are often treated as a single mechanism with one timing variable. What follows is a working repertoire of fourteen structures, each with characteristic purchases and costs. They are not mutually exclusive; most long novels use four or five, and the more interesting books use them in patterned sequence.

Seven of the fourteen require discussion beyond the table.

1. The alternating braid. Regular alternation is the safest structure and the one most likely to become invisible, which is both its virtue and its risk. Its underrated property is that regularity becomes a formal contract in the sense of Chapter 9 (The Opening Pages) §4.6: once a reader has learned that this book alternates, a departure from the pattern is legible as emphasis. Novels that establish a strict braid and then break it at the climax — running four consecutive chapters in one thread — are using the pattern as a rhythmic instrument, and the effect is unavailable to a book that never established one.

2. The block structure. The block is the highest-risk structure in the table and the source of the most common large-scale reading complaint, which is not confusion but reluctance. A reader two hundred pages into one thread does not want to leave it, and the transition is felt as a demand rather than as a variation. The books that succeed with blocks nearly always compensate by making each block generically distinct — a different voice, a different period, a different apparatus — so that the change reads as a new movement rather than as an interruption.

5. The repeated intersection. This is the characteristic structure of the great multi-plot novels and the least discussed in craft literature, because it has no climactic shape. Threads touch, affect each other by a degree, and separate: Lydgate treats Casaubon; Dorothea visits Rosamond; Farebrother carries a message. No single intersection resolves anything. The cumulative effect is a world in which everything is connected weakly rather than a plot in which two things are connected strongly, and that is a claim about society, not just a technique. Eliot’s web (§3.8) is the theory of it.

7. Thematic-only convergence. The structure usually treated as merely permissible, and treated here as one of the tradition’s summits. Its requirements are strict: the analogy must be marked often enough to survive the intervals (§6.6); the threads must share enough world for the comparison to be legitimate rather than authorial; and the ending must place them in relation without forcing a meeting. §9.2 gives the definitive case.

9. Promotion. The structure most likely to be mistaken for an error, because it is indistinguishable in the middle of the book from a hijack (§8.3). The difference is entirely one of design: in a promotion the transfer of primacy is the book’s argument, and the earlier main plot is revealed as having been a misunderstanding — Pip’s expectations were about Magwitch all along, and the reader’s misplaced investment is the novel’s subject. In a hijack the transfer is an accident of the writer’s interest, and the book’s centre of gravity moves without the book noticing.

13. The dropped thread. Almost always a failure, and included because the exception is so instructive. Melville introduces Bulkington in a page and then explicitly refuses to develop him, calling the short chapter his stoneless grave — a dropping that is announced, mourned and thematically loaded. The rule the exception yields: a thread may be dropped if the dropping is a narrative event. A thread that merely stops has not been dropped; it has been forgotten.

14. The digression proper. The digression has no chain, no goal and no convergence, and therefore fails the chapter’s definition at §2.1 outright. It is included because the greatest novels in the tradition contain a great deal of it, and because a chapter that could not say why Les Misérables is not ruined by Waterloo would have described a convention rather than a mechanism. §8.11 and §9.6 carry the argument.

6.5 How to use the taxonomy: substitution testing

The taxonomy’s purpose is not classification but substitution. A subplot that is not working is very often the right material in the wrong structure, and the fastest diagnostic is to ask what the same thread would be under a different arrangement.

Four substitutions worth running on any troubled thread:

The exercise is cheap because it is done on an outline rather than on prose. Re-chart the manuscript’s thread allocation on a single page — one row per chapter, one column per thread, a mark where each is active — and then redraw the marks under a different structure. Most structural problems are visible on that chart within a minute, and invisible in the manuscript for months.

6.6 Braiding rhythm: dormancy and reinstatement

A subplot braided in once or twice and then absent for two hundred pages cannot cash in a climactic convergence. That is a commonplace of craft advice, and the cognitive literature (§3.7) explains why, but the formulation leaves the writer without a working quantity. This subsection supplies one, with an explicit warning about its status.

The mechanism

When a narrative leaves a thread, the reader does not delete it. Emmott’s account (§3.7) is that the contextual frame is primed — held in a recoverable state — and reactivated on return. Reinstatement is therefore a recall operation, and recall has three properties the writer can act on:

1. It decays with interval. The longer the gap, the more cues are required.
2. It is cheaper for distinctive frames. A thread with an individuated cast, setting and register is recalled from fewer cues.
3. It is cheaper when the suspension left a question open. An unresolved question is an active memory item; a thread left at rest must be reconstructed from scratch.

The third property is the writer’s most powerful lever and the one most often wasted. Leaving a thread mid-question costs nothing and sharply reduces the reinstatement bill. (The size of the reduction is an inference from the frame research, not a measurement.) This is the braided form of the deferred decision from Chapter 2 (Scene and Sequel) and the proposition–payoff mechanism from Chapter 6 (The Chapter Break).

Working intervals

The table below states conventions derived from practice and calibrated against the cognitive literature. They are not experimental findings. No study has established how many pages a novel’s secondary thread may lie dormant, and any source claiming a number should be distrusted, this one included. What the table offers is a scale on which a manuscript’s behaviour can be described.

Three refinements.

The interval that matters is the longest one, not the average. A thread appearing in chapters 3, 5, 7 and 34 has an average interval that looks healthy and a single gap that destroys it. The audit at §5.5, step 5 measures the maximum for this reason.

Distinctiveness buys interval. A thread whose cast, place and register are strongly individuated can survive roughly one band longer than the table suggests. This is the operational form of §5.4’s second variable, and it is the reason the standard advice to differentiate secondary threads is correct for a reason rarely given: not for variety, but because distinctiveness is a memory subsidy.

A convergence must be paid into, and the payments must be recent. The rule of thumb the chapter offers, and marks as a rule of thumb: a thread that converges at the climax should have appeared in the quarter of the book immediately preceding it. A convergence prepared entirely in the first half and executed in the last chapter is a promise the reader has stopped holding.

The rhythm of the crossings themselves

Beyond dormancy, the pattern of crossings carries meaning, and four patterns can be named.

Accelerating is the default for a book with a literal convergence, and it is a technique writers rediscover rather than plan. It works because increasing crossing frequency raises the reader’s expectation of contact without any statement of it; the form is doing the foreshadowing.

6.7 Historical drift, and a caution against universalising the present

The conventions governing subplots have moved further than those governing any other unit in this textbook, and a student should be able to historicise the advice rather than receive it as timeless.

Two inferences follow, and they pull against each other, which is the point.

The advice in §§4, 7 and 8 is calibrated to the contemporary literary and commercial novel, whose length is a market fact (§3.6) and whose preference for tight causal linkage is a period style rather than a discovery. Applied to a nineteenth-century novel it produces confident errors; applied to a saga or a television season it produces the wrong arithmetic entirely.

The historical structures that violate it are not tolerated exceptions. Bleak House would fail a modern structural edit on at least four counts — cast size, thread count, coincidence, digressive social commentary — and it is widely ranked among the most formally accomplished novels in English. A pedagogy that can only say “an editor would cut this today” has described a market and not a book.

The double position to hold. Learn the present conventions well enough to break them deliberately; read historically enough to know that they are conventions, and that the form’s greatest achievements were built when they did not hold.