The price model carries an assumption that should be stated rather than smuggled in: that the reader might leave, that the book would rather she did not, and that attention is scarce. The assumption descends from the instalment market in which the next number had to be bought, and it describes one literary culture, broadly the trade fiction of the last half-century, not narrative as such. The method is useful in revision because most drafts are written within that culture.
It cannot describe several of the books this chapter has used. Tristram Shandy places its breaks arbitrarily, tears out a chapter and leaves others blank, in an attack on the idea that a book should carry its reader helplessly forward; its cuts are wrong by every test and right for a novel about the impossibility of ordering a life into units. A chapter read aloud at bedtime is designed to let the listener go, and children’s books are full of soft landings built to satisfy one reader and release another. The zhanghui chapter’s formulaic invitation to the next instalment descends from oral performance and owes nothing to the paperback. None of these is a fault the tests failed to catch, and none can be absorbed by widening their thresholds; they are segmentations whose value lies where the model sees only cost or waste.
The model therefore survives as an instrument for making a decision visible, so that a writer can confirm it as hers. A chapter that fails four tests may be the best in the book; what the method asks is that the writer know what each break charges, and why, and be able to say so. When the tests begin to choose for her, they have gone further than their evidence.
See also: the Reference coda; the Reference reading of Tristram Shandy.