12.1 Practical exercises
Exercise 1 — The eighteenth-century contents page. Write a summarising heading, in the manner of Fielding or Thackeray, for every chapter of your draft. *In which ___.* Produce the whole list on one page. Chapters that resist the formula have no proposition (§6.1, §7.1 test 1). Write 200 words on what the list revealed.
Exercise 2 — Six endings for one chapter. Take a single chapter and write six different final paragraphs, one for each type in §4: cliffhanger, soft landing, pivot, echo, question, recasting. Do not judge until all six exist. Then write 400 words on what each version does to the chapter that follows.
Exercise 3 — The substitution test. Choose the three chapter endings in your draft that feel weakest. For each, move the break one beat earlier and one beat later without rewriting a sentence (§6.5). Record all three versions and state which you keep and why.
Exercise 4 — The stopped reader. Give a reader four consecutive chapters of your draft and ask them to stop at each break, put the manuscript down for ten minutes, and write one sentence: what I want to happen next. Collect the four sentences verbatim. Do not explain, correct or discuss until all four are written.
Exercise 5 — The chapter-length audit, on a published novel. Count the words in every chapter of a novel you know well. Plot them in order. Write 500 words on the shape of the curve, what it corresponds to in the novel’s action, and whether the mean or the variance is the more informative statistic (§5.4).
Exercise 6 — The four-column table. Produce the audit at §5.6 for your own manuscript. Bring the table, not the manuscript, to the seminar, and be prepared to have the book diagnosed from it.
Exercise 7 — Re-segmenting. Take twenty pages of a novel that uses long chapters and re-break them into six short ones. Then take twenty pages of a thriller and merge them into two. Read both aloud. Write 300 words on what changed when nothing changed.
Exercise 8 — The instalment rewrite. Take four consecutive chapters of your draft and re-plan them as a single monthly number in the Victorian manner: decide which effects fall where, and which chapter has to carry the ending that survives a month. Write the number plan as Dickens wrote his — a list of topics, characters and deferrals — and then say what the exercise changed about your ordering (§3.3, §9.1).
Exercise 9 — The unnarrated chapter. Identify the event in your draft that matters most and is currently narrated. Remove it, place a chapter break where it would have been, and rewrite the chapters on either side so that the event is present only in its approach and its consequence (§6.4, strategy 12). Write 400 words on whether the novel survived, and on what would have to be true for it to survive.
12.2 Seminar questions
Is the chapter a technology or a form? Dames argues that the chapter is an inherited technology of segmentation rather than a natural unit of narrative. If that is right, does it follow that any placement is as good as any other, or does the argument leave the craft of placement untouched? Test the claim against Tristram Shandy and against a contemporary thriller.
Five ending types or six? Argue for or against admitting the recasting close (§4.6) as a category distinct from the pivot. Use Things Fall Apart Chapter 7, The Great Gatsby Chapter 3, and Vanity Fair Chapter 32 as evidence. Note that all three are arguably both.
Was the cliffhanger a commercial device that became a craft, or a craft device that the market happened to reward? The Count of Monte Cristo and Bleak House were both shaped by delivery formats with real intervals. Does knowing this change how the endings should be evaluated?
Whose conventions? The failure patterns in §8 are drawn from Anglo-American trade practice since roughly 1960. Examine the segmentation of a novel from another tradition — the zhanghui chapters of classical Chinese vernacular fiction, or the divisions of The Tale of Genji, or a novel translated from a language whose publishing conventions differ — and identify which of §8’s prohibitions describe narrative and which describe a local literary culture.
Does the digital exit change the craft? Streaming has removed the act of stopping from television; the e-reader has made the chapter a displayed quantity of time. If the reader no longer has to decide to continue, does the chapter ending still have work to do — and is the answer different for literary and commercial fiction?
Should a novel be readable in any order? Catch-22 and Moby-Dick both use segmentation to weaken sequence; Hopscotch and The Dictionary of the Khazars go further. Is a chapter’s position in a sequence a property of the chapter, or only of the book? What would a novel lose if its chapters could be read in any order?
12.3 Assessment rubric
For a submitted sequence of three consecutive chapters (4,000–7,000 words total) with a 1,000-word critical commentary.
| Band | Craft (60%) | Commentary (40%) |
|---|---|---|
| Distinction | Each chapter has a statable proposition and an identifiable payoff; the three endings differ in type and in price, and the differences are motivated by the sequence; lengths are proportionate to weight; at least one break is placed somewhere other than the obvious joint, defensibly | Names each ending type and placement strategy and defends them against the alternatives it tried; identifies its own risks; engages Dames or the serial history critically rather than decoratively; assesses what the sequence does that no single chapter does |
| Merit | Propositions present; payoffs identifiable; some variety of ending type; lengths broadly proportionate; placements defensible if conventional | Accurate self-analysis; substitution alternatives considered; some engagement with theory; risks acknowledged if not resolved |
| Pass | Chapters are complete units; endings are competent; variety limited; placements fall at the obvious joints throughout | Describes the endings accurately; theory referenced but not applied; no evidence of alternatives having been tried |
| Fail | No statable proposition; endings arbitrary or uniformly cliffhung; length uniform by default; one or more failure patterns unrecognised | Description without analysis; no engagement with the model; the sequence is not treated as an object |