1. Introduction: The Scenery Problem

1.1 An asymmetry between writing and reading

Consider two passages from the same manuscript, each about nine hundred words long.

In the first, two sisters clear out their dead father’s shop. They talk. One of them makes a joke about the state of the stockroom; the other does not laugh. They find a box of receipts, decide it can wait, and go to get lunch. The prose is exact. The dialogue is plausible. The reader finishes the passage having enjoyed it in the mild way one enjoys competent observation, and remembers, a day later, nothing.

In the second, the same two sisters clear out the same shop, and one of them finds a page in the ledger that is not in their father’s handwriting. She says nothing about it. She goes on wrapping glassware, and when her sister asks what she has found, she says “receipts”. The prose is no better. The dialogue is no more plausible. But the reader now knows something the second sister does not, is left with a question the book has not answered, and will read the next forty pages differently.

Two consequences follow, and they organise everything that comes after.

• The diagnosis has to happen in revision. The tests in §7 are useless and actively harmful during composition. They are for the second pass, when the manuscript has become an object rather than a process.
• Prose quality is not evidence. A well-written scene that does nothing is harder to cut than a badly written one, and is cut less often, which is why manuscripts accumulate them. The most expensive material in a draft is the material the writer is proud of.

1.2 The claim in its strong form

The principle can be put in twelve words:

If nothing changes, the scene is not a scene — it is scenery.

The formulation is deliberately harsh, and the rest of this chapter is largely a matter of establishing what it does and does not claim.

• It does claim that a stretch of dramatised narrative is answerable for its consumption of the reader’s attention and the book’s length, and that the currency in which it answers is change of state.
• It does not claim that the change must be large, external, or visible to the characters. The change in The Portrait of a Lady, chapter 42, is that a woman sitting motionless by a dying fire understands her marriage; nothing else in the room alters (§9.8).
• It does not claim that summary is a lesser mode. A great deal of what a draft calls a “flat scene” is a paragraph of summary that has been inflated into dialogue and blocking because the writer believed that scenes are always better. They are not (§6).
• It does not claim universality. The preference for the dramatised scene over the narrated report is a local and recent convention of Anglo-American fiction and its workshop pedagogy, and eighteenth- and nineteenth-century practice differs so widely that the model has to be historicised rather than applied (§3.6, §6.5).

1.3 Where this chapter sits in the series

This chapter is the foundation of the textbook, in the ordinary sense that later chapters depend on its vocabulary.

Chapter 2 (Scene and Sequel) owns the rhythm. Dwight Swain’s division of the dramatic unit into a scene proper (goal–conflict–disaster) and a sequel (reaction–dilemma–decision) is treated there at length, including its origins in mid-century American commercial-fiction pedagogy and the reasons to be sceptical of it. The present chapter uses “scene” in the loose, ordinary sense — any continuous stretch of dramatised action — and refers to Swain’s narrower sense only where the distinction bears on diagnosis (§6.6).