Only the theory the chapter’s method uses is set out here, and each strand begins with the practical question it answers. The full treatment gives each at greater length.
Does the middle have a job of its own?
A writer who thinks of the middle as the distance between a beginning and an end will fill it, and the question is whether it is anything more. The oldest answer is Aristotle’s. The Poetics says that a whole has a beginning, a middle and an end, and defines the middle as “that which follows something as some other thing follows it” (in S. H. Butcher’s translation, 1895). The definition is causal, not positional: a middle is identified by what it follows from and what follows from it, so a stretch of action that neither depends on what came before nor produces what comes after is not a middle on Aristotle’s terms, however central its pages. The same treatise draws the consequence. “Of all plots and actions the episodic are the worst,” it says, and an episodic plot is one “in which the episodes or acts succeed one another without probable or necessary sequence.” The test that follows, whether two parts of the middle could change places, is the reorder test in Diagnostic Tests.
The second answer comes from the theatre. Gustav Freytag’s Die Technik des Dramas (1863), translated by Elias J. MacEwan as Technique of the Drama in the 1890s, describes a five-act tragedy as a pyramid in five parts: introduction, rise, climax, return or fall, and catastrophe. The climax stands at the centre, in the third act, and everything after it is consequence. Most modern novels put their climax late, so the pyramid fits them badly, but it records something the later models keep: the centre of a dramatic action is where its direction is decided, not where it waits. The midpoint of the modern manuals is what remains of Freytag’s central climax in a form that has moved its climax to the end.
Why will a reader put up with waiting?
The middle withholds what the reader wants to know, for most of the book, and the writer’s question is why a reader tolerates it and when the tolerance runs out. Roland Barthes’s S/Z (1970), translated by Richard Miller (1974), reads a Balzac story through five codes, one of which, the hermeneutic code, is the set of devices by which a text raises a question and delays its answer. Barthes names the delays: the snare, which points the reader the wrong way; the equivocation, which mixes truth and snare; the partial answer; the suspended answer, broken off once begun; and jamming, the admission that the question cannot be answered. Peter Brooks’s Reading for the Plot (1984) builds on this to describe the middle as a dilatory space, the stretch in which the end is postponed, and in a reading of Freud argues that its detours are necessary: without them the action reaches its end too soon, by a short circuit, and arrives at the wrong end, one the book has not prepared.
The two accounts give the craft rule that the chapter uses throughout. Delay is the middle’s medium, and the question is never whether to delay but how. Barthes’s delays differ in what they do to the reader’s question. A partial answer changes it, since the reader now knows something and wants something else, and a snare changes it by sending the reader after the wrong thing. A delay of the same kind repeated, the question put and blocked and put again, leaves it where it was, and that is the delay readers experience as stalling. The household draft in The Problem delays by the same device five times.
Where should the turns fall?
Writers want to know where in the middle its turns belong, and the most detailed answers come from screenwriting manuals. Syd Field’s Screenplay (1979) divided a feature film into three acts, with Act Two, which he called the confrontation, filling the middle half. In The Screenwriter’s Workbook (1984) he added a midpoint, which he said he had found while studying Robert Towne’s screenplay for Chinatown, and his later accounts of the paradigm add a smaller turn in each half of Act Two, at about pages 45 and 75, which he called a pinch, a scene that keeps the story on track; the manuals after him made it a reminder of the opposing force. Blake Snyder’s Save the Cat! (2005) made the midpoint a false victory or a false defeat, a point at which the protagonist seems to have won or lost and the stakes are raised, after which the opposition closes in. James Scott Bell’s Write Your Novel From the Middle (2014), written for novelists, calls the centre the mirror moment: the protagonist takes a hard look at what they have become, or realises that they are likely to die.
The manuals converge on a turn near the centre that changes the protagonist’s relation to the goal, and that convergence is the reason this chapter uses the midpoint. Their placements are conventions of one commercial form whose running time is fixed, and they describe current practice rather than laws. The novel’s middles were shaped earlier by its forms of publication. A three-volume novel invited a turn inside its middle volume: in Pride and Prejudice (1813) Darcy’s first proposal falls in the eleventh of the second volume’s nineteen chapters. A novel published in monthly parts, as Dickens’s were, needed a turn in every part, so its middle was a chain of small turns rather than one central one. The tests in this chapter therefore ask what a middle does, whether it narrows the book, and not where its turns sit.
Is there evidence?
The evidence is thin. A study by Andrew Reagan and colleagues, published in EPJ Data Science in 2016, measured the emotional vocabulary across some 1,700 works of fiction from Project Gutenberg and found that their trajectories cluster into six basic shapes. The two shapes with two changes of direction, a rise, fall and rise and a fall, rise and fall, were associated with the most downloads. The measure is of words, not of events, and downloads are not a measure of reading, so the finding is consistent with the claim that readers prefer a middle that turns and does not establish it. This chapter treats that claim as a conjecture supported by editorial experience.
See also: Scene and Sequel (Chapter 2), where Brooks’s account of narrative desire explains why a single scene should end in a setback; The Ending (Chapter 16), for Brooks on the retroactive reading of the whole, which is the other half of the dilatory space.