10. The Subplot Across Forms

The domain model and the attention tax were developed for the novel. They transfer to adjacent forms with adjustments that should be stated, because writers routinely import subplot habits into forms that punish them.

10.1 The short story

The short story is the form in which the subplot is nearly always wrong, and the reason is arithmetical rather than aesthetic.

What stories use instead is the compressed analogy: a second situation glimpsed rather than narrated — a neighbour’s marriage overheard, a parallel case mentioned once, a story told inside the story. It supplies Empson’s juxtaposition at a cost of a paragraph. The nested subplot (§6.4, structure 11) is the only structure from the taxonomy that transfers to the story intact, and it transfers because nesting requires no braiding.

The exception is the linked collection, in which each story is a thread and the book is the braid. There the relations are Levin’s — analogy and contrast, almost never direct linkage — and the convergence, when it occurs, is thematic by necessity.

10.2 The novella

The novella’s characteristic subplot problem is that it is written by novelists. It has a novel’s tonal amplitude and a story’s page count, and a thread that would be modest in a 120,000-word book occupies a quarter of a 30,000-word one.

The form’s usual solution is the single-scene subplot: a secondary chain whose entire existence is one scene early and one scene late, with the middle inferred. The Death of Ivan Ilyich uses a version of it — the colleagues’ reaction to the death at the beginning, and the household’s arrangements at the end, frame a life told with almost no second chain at all.

The general rule: in the novella, prefer a second chain that is two scenes long over one that is six scenes long. The reader will construct the intervening chain, and the construction costs the book nothing.

10.3 Series and sequences

The series introduces a structure the single novel does not have: a thread that is a subplot in one volume and the main plot in another. This is promotion (§6.4, structure 9) distributed across books, and it is the form’s chief structural pleasure.

Three principles:

1. A volume must be legible as a whole. A thread that only makes sense in book four has failed inside book two, whatever its eventual payoff. The convergence test applies per volume as well as across the sequence.
2. Dormancy intervals should be calculated across the sequence, not within a volume. A thread absent for a whole book has been dormant for a year of the reader’s life, and §6.6’s final band applies with force. Series writers systematically underestimate this.
3. Promotion must be prepared in the volume before it happens, not merely in the volume where it began. The reader’s memory is of the last book they read.

Chapter 17 (Series and Genre) treats the series as a form, and the dormancy of threads across volumes.

10.4 The screen comparison, and what does not transfer

Long-form television is the current home of braided narrative, and its vocabulary — A-plot, B-plot, C-plot, the season arc, the “runner” — has entered prose craft discussion wholesale. Two features transfer and three do not.

What transfers.

• Per-unit resolution. The discipline of giving each episode’s B-plot a small completed shape is a good discipline for chapters, and Chapter 6 (The Chapter Break) makes the case.
• The accelerating braid. Television’s tightening of cross-cutting towards a finale is the pattern §6.6 recommends, arrived at independently.

What does not.

• Involuntary reinstatement. A returning thread on screen re-establishes itself the instant a face appears. Prose has no faces; every resumption is paid for in words. This is the single largest difference between the media and the reason television can sustain dormancy intervals that would destroy a novel.
• Actor-driven page-share. Television allocates screen time partly for contractual and production reasons, which produces the occupational subplot (§8.8) as an industrial by-product. Novels have no such excuse and should not imitate the result.
• The obligation to continue. A season must generate next season’s threads, which produces deliberate non-convergence. A novel that leaves threads open for a series instalment it may never write has borrowed a structural habit from an economy it does not share.

10.5 Translated fiction

Four of this chapter’s eight close readings were written in Russian or French, and the reliance is not accidental: the multi-plot novel’s greatest achievements are largely continental, and an anglophone student meets them in a translator’s English.

The distinction this book’s Introduction draws between structural and textural claims applies with unusual clarity here, and to the chapter’s advantage. Plot architecture is the most translation-stable feature of a novel. Braiding intervals, page-share, convergence points, the relation between chains — all are properties of the story’s arrangement, and all survive intact into any competent translation. A student can argue about the structure of Anna Karenina from Garnett, from the Maudes, from Pevear and Volokhonsky or from Bartlett, and the argument will be the same.

Two cautions:

• Abridgement is a structural intervention. Several nineteenth-century French and Russian novels circulate in English in abridged or “reader’s” editions from which the digressions have been removed. A student arguing about the proportion of Les Misérables occupied by Waterloo must first establish which text they are holding.
• Volume divisions and part titles are frequently editorial. Where an argument depends on where a book’s Parts fall — and several arguments in §6 do — the division should be checked against the original publication rather than against a paperback’s table of contents.

10.6 Concordance: concept to novel

A study aid. Each entry names the work that most clearly displays the concept.