The Craft Behind the Story Read online

Chapter 6 · The Chapter Break

6. Diagnostic Tests

Continue in Reference

The ten tests are for revision, not composition; applied while drafting they produce chapters shaped like the tests. Tests 1 to 6 (Proposition, Payoff, Permission, Stopped reader, Placement and Proportion) are applied to one chapter and tests 7 to 10 (Variety, Re-entry, Container and Architecture) to the manuscript, because much of what goes wrong at chapter breaks is visible only in the aggregate. Run tests 1 (Proposition) and 2 (Payoff) first, since there is no point pricing the exit of a chapter that proposes nothing. Record a sentence, not a verdict: “stopping here is free, and this is the fourth free stop in a row” is a revision plan, and “fails test 3” is not. Each test is failed on purpose by a book worth reading.

Test 1 — Proposition

Can I state in one sentence the question this chapter raised near its start? Can I write its heading, In which…?

Failure. No subject: a stretch of narrative that happens to be bounded, or atmosphere, which cannot be paid off.

Exception. Tristram Shandy, whose chapters mock the idea of a subject.


Test 2 — Payoff

Does the ending resolve, escalate, hold open, recast or exchange that proposition?

Failure. Abandonment: the chapter drifted from its opening.

Exception. The Trial, whose questions are never paid and accumulate into the subject.


Test 3 — Permission

Is stopping here free, moderate or expensive, and why that price at this point?

Failure. The price was not chosen.

Exception. A daily feuilleton, whose price the format set.


Test 4 — Stopped reader

Close the file for ten minutes, then write what happened, what is open, what you want and how you feel.

Failure. Nothing wanted; a want attached to nobody; events and no feeling.

Exception. Middlemarch, whose reader leaves chapters holding an idea.


Test 5 — Placement

Moved one beat earlier and one later, can I say why the present position is right?

Failure. A defensible cut that is not the best one.

Exception. Don Quixote I.8, a cut visibly in the wrong place on purpose.


Test 6 — Proportion

Does this chapter’s length match its weight in the book?

Failure. Padding towards a habitual length, or a turning point squeezed.

Exception. Moby-Dick, whose digressions are given chapters of their own.


Test 7 — Variety

Across eleven consecutive chapters, are the ending types and lengths chosen or defaulted to?

Failure. One gear: all cliffhangers, or all soft landings.

Exception. A thriller whose single gear is its contract with the reader (The Da Vinci Code).


Test 8 — Re-entry

How many words does the next chapter spend before its own business begins?

Failure. Over about sixty words, as a rule of thumb: the break was placed where nothing changed.

Exception. A serial instalment, whose reader has been away.


Test 9 — Container

Is this one chapter, or two not yet separated, or one cut in half?

Failure. Test 2 (Payoff) fails for the whole and passes for each half.

Exception. A Middlemarch chapter, held together by its Book.


Test 10 — Architecture

Reading only the headings, Part titles and first and last sentences, is the book’s shape legible?

Failure. Stretches of interchangeable endings; Parts where nothing changes.

Exception. One Hundred Years of Solitude, with no numbers, titles or Parts.

Three refinements sharpen the tests. In test 2 (Payoff), read a chapter’s first page and its last paragraph consecutively, skipping the middle: a well-shaped chapter survives the treatment, and one that does not has drifted or is two chapters. Test 5 (Placement) is slow, about twenty minutes a chapter, so reserve it for the breaks that carry the most weight, the Part boundaries, the viewpoint handoffs and the ends of movements; its common result, that all three positions work, frees the writer to place the break for the sequence. Test 10 (Architecture) requires a compiled document and an hour’s work, and asks whether the sequence of last sentences has a shape.

Reading the results together

Failures cluster, and the cluster points to the first move. Tests 1 (Proposition) and 2 (Payoff) together mean the chapter has no subject: do not touch the ending, but find the proposition or merge the chapter with a neighbour. Test 2 (Payoff) alone means abandonment: return to the opening question on the last page, or move the break to where the first proposition completes. Test 4 (Stopped reader) alone means an informational ending: change the last paragraph, usually to an echo or recasting close. Test 5 (Placement) means a cut that should move, which costs no prose. Tests 3 (Permission) and 6 (Proportion) together mean a chapter sized and ended by habit, which suggests a manuscript-wide pattern; go to test 7 (Variety). Test 8 (Re-entry) means orientation is being paid on the wrong side of the break: move it to a decision or an announced destination. Test 9 (Container) means split at the internal seam or merge and re-cut.

Test 7 (Variety) carries a warning. Writers who find their endings monotonous tend to respond by adding cliffhangers, because the cliffhanger feels like an improvement. A manuscript with too many expensive exits fails the same test as one with too few, and in ambitious drafts, by this book’s estimate from manuscripts read rather than from a count, too many is the commoner condition, so the corrective is usually to install soft landings at the ends of movements.

See also: the Reference protocol, with each test’s procedure and a table of combined failures; the Reference failure catalogue.