The Middle

Narrowing, Delay and the Long Stretch: The Full Treatment

The Craft Behind the Story (Second Edition), Reference — Chapter 11. Part III: The Architecture of the Whole.

This chapter is new in the second edition, so its Reference layer is this single file rather than a corrected original. It extends the reader’s version in Core, The Middle, and can be read without it: the definitions it relies on are restated in section 1. Illustrative passages are quoted briefly for the purposes of criticism and study, and full citations appear in section 9. Where a work is read in translation, the translator is named.

Contents

1. Scope and working definitions
2. The theory at length: Aristotle and Freytag; Barthes’s delays; Brooks’s detour; the screenwriting conventions compared; forms of publication; the evidence
3. The extended catalogue of shapes, with three further shapes and a summary table
4. The extended catalogue of failure patterns
5. The diagnostic procedures
6. Further analyses: Catching Fire, The Seven Husbands of Evelyn Hugo, the second stage of Great Expectations, Squid Game
7. The household middle in full
8. The limits of the model
9. Sources

1. Scope and working definitions

The middle of a novel is the stretch between the first commitment its beginning cannot take back and the start of its ending, the climax and its aftermath as Chapter 16 defines them. The first boundary is a change of state: the protagonist has entered the line of action the book will follow, and the opening’s normal is no longer available. The second is the start of the sequence that leads straight into the climax. Defined this way the middle is usually the largest part of a book, commonly half to three-fifths of it; in the household novel it runs from the register drawer at about 25 per cent to Dov’s chapter at about 78 per cent, 53 per cent of the book.

The chapter’s mechanism rests on a small vocabulary. A complication is anything that stands in the protagonist’s way. An escalation is a complication that changes what the protagonist can do or what doing it will cost. The option list is the set of moves open to the protagonist towards the goal at a given point; an event that removes a move narrows the list, an event that makes a move cost more raises its price, and an event that does neither only delays. The midpoint is a turn near the centre of the middle that changes the protagonist’s relation to the goal, by changing the goal, the protagonist’s knowledge, their mode (from reacting to acting) or their fortune. The value-shift log is a list of the middle’s sequences, each with the change it makes, written as two parallel sentences in the manner of Chapter 1’s value shift, and the stakes in a word.

The governing claim is that a middle sags when its events leave the option list and its prices unchanged, however many events there are, and that its work is to narrow what the book can still do until the ending is the move that remains. The claim is about function, not placement. A middle can narrow with its turns anywhere, and a turn at the centre that closes nothing does not satisfy it.

The chapter does not treat the escalation of a single scene-and-sequel pair (Chapter 2), the architecture of plot lines (Chapter 14), the protagonist’s inner change across the book (Chapter 12, Arc and Theme), the management of a revelation (Chapter 15, The Reveal), middles told out of order (Chapter 8, Non-linear and Multi-timeline Architecture), the middle volume of a series (Chapter 17, Series and Genre) or the order of revision (Chapter 18, The Structural Edit).

2. The theory at length

2.1 Aristotle and Freytag

Aristotle’s Poetics defines a whole as that which has a beginning, a middle and an end, and the middle as “that which follows something as some other thing follows it” (Butcher’s translation). The definition makes the middle a causal position: it depends on the beginning and produces the end. Two further passages bear on the middle. The ninth chapter calls the episodic plot, “in which the episodes or acts succeed one another without probable or necessary sequence”, the worst of all; and the eleventh describes the two parts of a plot that most move an audience, reversal (peripeteia), a change of the situation into its opposite, and recognition (anagnorisis), a change from ignorance to knowledge. Aristotle located both at the turning point of a tragedy, which in a short play is near its end. In the long novel they migrate: Elizabeth Bennet’s recognition in Pride and Prejudice, Till this moment I never knew myself, sits at the centre of the book, and the modern manuals’ midpoint is in large part the old reversal and recognition moved to the middle of a longer form.

Gustav Freytag’s Die Technik des Dramas (1863), translated by Elias J. MacEwan in the 1890s, describes the five-act tragedy as a pyramid of five parts, introduction, rise, climax, return or fall, and catastrophe, with the climax at the centre. The pyramid fits the modern novel badly, since most novels place their climax late, and later diagrams that move the climax towards the end describe a different model under Freytag’s name. What the original keeps for this chapter is the claim that the centre of a dramatic action is where its direction is settled: everything after Freytag’s climax is consequence. A novel’s midpoint does the same work at lower intensity.

2.2 Barthes’s delays

Roland Barthes’s S/Z (1970), translated by Richard Miller (1974), reads Balzac’s story “Sarrasine” through five codes. The hermeneutic code is the set of units by which a text raises an enigma and, after delays, answers it, and Barthes names the delays the code uses between question and answer. They are the middle’s tools, and they differ in what they do to the reader’s question, which is the difference this chapter needs. The table sets them out with a use from the household novel where one exists.

Read as a set, the delays show why some waiting is pleasant and some is not. Partial answers and snares change the question, so the reader is moving even while the answer is withheld. A suspended answer holds the question still but puts the answer in view, which is tolerable once and maddening if repeated. A delay that repeats the same blockage, the question put and refused and put again, is not in Barthes’s list at all, because it does nothing to the question. It is the plateau’s delay.

2.3 Brooks’s detour

Peter Brooks’s Reading for the Plot (1984) takes Barthes’s account into a theory of narrative desire. Brooks calls the middle the dilatory space of the text, and in his chapter on Freud’s Beyond the Pleasure Principle he argues that the middle’s detours protect the end: an action that reached its end at once, by a short circuit, would arrive at the wrong end, one the text had not earned. The middle exists so that the end can be the right one. For a writer the argument has two consequences. Delay is not a failure of the middle but its medium; and the detour must be a detour towards the end, not away from it, which is the narrowing claim in another vocabulary. Brooks’s account of reading as anticipation of retrospection, in which the reader moves forward expecting the end to confer meaning, is the other half of the dilatory space and is treated in Chapter 16.

2.4 The screenwriting conventions compared

The most detailed advice on the shape of a middle comes from screenwriting manuals written for a form whose running time is fixed. The table compares three that are widely used, with their placements converted to percentages of the script.

The three agree on a turn near the centre that changes the protagonist’s relation to the goal, and they agree on little else. Field’s pinch and Snyder’s closing-in describe what this chapter calls the opposition’s chain surfacing as consequence. Snyder’s false victory and false defeat are reversals of fortune; Bell’s mirror moment is a change of mode or of self-knowledge. None of them distinguishes narrowing from price, which is where most middles fail.

2.5 Forms of publication

Before the screen manuals, the shape of a novel’s middle was set largely by how it reached readers, and the conventions above are recent. The three-volume novel of the nineteenth century gave its middle volume a natural turn: Pride and Prejudice (1813) has volumes of 23, 19 and 19 chapters, and Darcy’s first proposal falls in the eleventh chapter of the second. A novel issued in monthly parts, as most of Dickens’s were, needed a turn in each part, so its middle was a chain of small turns rather than one large central one, and such novels often read as a series of narrowings without a single midpoint. Great Expectations (1860–61), serialised weekly, marks its own structure instead, in three named stages (section 6.3). Television’s season structure has produced a further convention, the turn at about two-thirds of a season (section 6.4). The diagnostic tests ask what a middle does rather than where its turns sit, because placement conventions are period-bound and function is not.

2.6 The evidence

The empirical evidence is thin and indirect. Andrew Reagan and colleagues, in EPJ Data Science (2016), measured the emotional vocabulary of some 1,700 works of fiction from Project Gutenberg along their length and found that the trajectories cluster into six basic shapes: a steady rise, a steady fall, a fall then a rise, a rise then a fall, and two shapes with two changes of direction, rise–fall–rise and fall–rise–fall. The two-turn shapes were associated with the most downloads. The study measures words rather than events, and downloads rather than reading, so it is consistent with the claim that readers prefer a middle with a turn in it and does not show it. The chapter’s claims about narrowing are conjectures from editorial practice, offered as testable in a writer’s own draft.

3. The extended catalogue of shapes

Middles are classified here by the engine that carries their pressure, that is, by the means by which they narrow the book. Six shapes are treated in the Core chapter and summarised here; three further shapes follow, and a table closes the section.

3.1 The six shapes in brief

The escalating chain carries its middle on one line of action whose sequences cause each other and cost more each time; Highsmith’s The Talented Mr Ripley (1955) is the pure case, each of Tom Ripley’s acts demanded by the last. The turned middle divides at a central turn into two halves that ask different questions; Catching Fire (section 6.1) and Pride and Prejudice are its cases. The stations middle joins encounters by a constant, a road, a list, a frame, and must import its narrowing from a rule or a limit; the picaresque, Don Quixote and The Midnight Library are its cases, and The Seven Husbands of Evelyn Hugo (section 6.2) is a stations middle held by a frame. The braided middle alternates plot lines and narrows by their crossings, as Chapter 14 describes. The accumulating middle gathers knowledge through Barthes’s delays; Collins’s The Moonstone (1868), whose successive witnesses each close some explanations of the theft, is its case. The contained middle holds its protagonist in a fixed situation and takes its engine from time and arrivals; Mann’s The Magic Mountain (1924), translated by John E. Woods (1995), and Towles’s A Gentleman in Moscow (2016) are its cases.

3.2 The mirrored middle

A mirrored middle is built symmetrically around its centre, so that events of the first half are answered, often reversed, by events of the second, and the centre is the axis. Fielding’s Tom Jones (1749) is the classic case: its eighteen books fall into three groups of six, the country, the road and London, and the comedy at the inn at Upton in Books 9 and 10 sits at the centre, where the novel’s lines of pursuit cross and change direction. The engine is anticipation of the answering half: once a reader has felt the symmetry, each event of the second half is read against its partner in the first, and the narrowing is the closing of the pattern. The characteristic failure is symmetry that is visible to the writer and not to the reader, which produces a second half that repeats the first without the reader knowing why.

3.3 The interval middle

An interval middle is a gap. The book’s centre is a span of time passed over, briefly told, between two blocks of scene, and the narrowing happens in the gap: what was possible before it is no longer possible after. Virginia Woolf’s To the Lighthouse (1927) puts ten years and several deaths into its short central section, “Time Passes”, and Shakespeare’s The Winter’s Tale puts sixteen years into a single speech by Time. The engine is consequence revealed at once: the reader returns to the action to find the option list already narrowed by events summarised or omitted. The characteristic failure is a gap that changes nothing, a time-jump used to skip a middle the writer could not write, after which the second block resumes the first.

3.4 The retold middle

A retold middle turns at its centre by changing vantage: the second half tells again, from another position, what the first half told. Lauren Groff’s Fates and Furies (2015) gives the first half of a marriage’s story, “Fates”, largely to the husband, Lotto, and the second, “Furies”, to the wife, Mathilde, so that the second half narrows the first by revealing what its teller did not know. The engine is knowledge, as in the accumulating middle, but delivered as a single block rather than by instalments, which makes the retold middle a turned middle whose turn is a change of teller. The characteristic failure is a second telling that adds texture without revising anything the reader believed, so that the reader rereads without being repaid. Multi-vantage structures in general belong to Chapter 13.

3.5 Summary table

The table gathers the nine shapes with their engines and their characteristic failures, as a quick reference for diagnosis.

Most books combine two or three of these, and the useful question in revision is which engine is carrying the pressure in each half of the middle. The household novel’s middle is braided and accumulating, and its revision (section 7) makes it braided and turned.

4. The extended catalogue of failure patterns

The Core chapter gives five patterns in full, each with a published exception; they are summarised here and followed by six more. For each additional pattern the entry gives symptom, mechanism and remedy.

4.1 The five patterns of the Core chapter

The plateau repeats one value shift at constant stakes, leaving the option list unchanged; its exception is Beckett’s Waiting for Godot, which Vivian Mercier called a play in which “nothing happens, twice”, and whose repetition degrades. The episodic middle joins its sequences by a constant rather than cause; its exception is Don Quixote, translated by Edith Grossman (2003), whose second part supplies a cause the first lacked. The absent opposition lets the opposing force vanish while the protagonist prepares; its exception is Moby-Dick (1851), whose whale appears only in chapter 133 of 135 but acts through the nine gams. The holding pattern waits for an event scheduled at its end; its exception is Tartt’s The Goldfinch (2013), whose Las Vegas stretch conceals the book’s central plant. The broken back turns so hard at the centre that the second half is another book; its exception is Hitchcock’s Psycho (1960), which kills its protagonist about 43 per cent of the way through and sinks the first half’s question in a swamp on screen.

4.2 The training-montage middle

Symptom. The protagonist spends the middle acquiring the skill, knowledge or equipment the climax will need: lessons, rehearsals, research, a journey to the place of confrontation.

Mechanism. Acquisition opens moves rather than closing them, so the option list grows through the middle and the ending must close everything at once. The reader also knows that the acquisition will succeed, since the climax requires it, so its sequences have no price.

Remedy. Make each acquisition cost a move: the skill learned at the price of an ally, the knowledge at the price of a secret kept. Or summarise the acquisition and spend the middle on what it costs.

4.3 The spectacle that turns nothing

Symptom. The middle’s largest set piece, a battle, a party, a storm, a death, sits at or near the centre, and the book after it continues as before.

Mechanism. The event is large and the option list is unchanged, so the reader registers size without consequence, which after the first time teaches them to discount size. It is a false midpoint: an event at the centre rather than a turn.

Remedy. Run the midpoint test on the set piece. If it changes nothing among goal, knowledge, mode and fortune, either give it a consequence that closes a move or move it, since a large event that changes nothing is better placed where it is not mistaken for the turn.

4.4 The reset

Symptom. Each sequence ends by restoring the situation it began with: the danger averted, the quarrel made up, the secret kept for another day.

Mechanism. The reset is the norm of episodic series television, in which each episode returns its characters to a stable state so that episodes can be watched in any order. In a novel it produces a middle of self-cancelling sequences, each of which leaves the option list as the previous sequence left it. The result is a plateau achieved by repair rather than repetition.

Remedy. Let something survive each sequence: a cost paid, a fact known, a relation changed. The reset is acceptable as an occasional rest, and a book can use one to show that a character cannot change.

4.5 The research middle

Symptom. The middle delivers what the writer learned: the history of the setting, the workings of the trade, the procedures of the institution, in sequences whose purpose is to inform.

Mechanism. Information arrives without pressure, in the terms of Chapter 1’s pressure test, so it answers questions the reader has not asked, which is the reverse of Barthes’s partial answer.

Remedy. Attach each piece of information to a move it opens or closes for the protagonist, and cut what cannot be attached. The exception is a book whose information is its pleasure, as the cetology chapters of Moby-Dick are for the readers who love them, where the information is the texture of the voice rather than a delay of the plot.

4.6 The second sag

Symptom. A strong midpoint is followed by a long stretch in which the second half’s question is set and nothing further closes until the climax.

Mechanism. The midpoint has done its narrowing at once, and the writer, relieved, lets the clock or the coming climax carry the rest. It is the commonest failure of middles that have already been revised once, because a first revision usually installs a midpoint and does not check what follows it.

Remedy. Run the option list at three-quarters as well as at the midpoint, and give the stretch between at least one sequence that closes a move, ideally by the opposition’s chain surfacing. The household revision has a stretch of this kind (section 7).

4.7 The premature climax

Symptom. The book’s central conflict is discharged in the middle, and the second half restarts it, often with a new antagonist or a raised stake.

Mechanism. Brooks’s short circuit: the action has reached an end too early, and the second half must manufacture a new middle to reach the real one. Readers feel the book has ended and begun again.

Remedy. Decide which climax is the book’s. If it is the first, the second half is an aftermath and should be shortened accordingly, in the terms of Chapter 16’s proportion problem; if it is the second, the first must be reduced to a turn that leaves the conflict open.

5. The diagnostic procedures

The Core chapter’s eight tests need a summary of the middle to work on: one line per sequence, in order, with its position as a percentage of the book. Four of them are procedures that repay a fuller description.

The value-shift log (test 1). For each sequence, write what is true at its start and what is true at its close, in two parallel sentences, and the stakes in one word. A blank log has four columns: position, sequence, shift, stakes. Then read down the shift column and mark each neighbouring pair whose shifts are the same kind of change, and down the stakes column for each pair with the same stake. A pair marked in both columns is a plateau, and a run of three is a plateau the reader has already noticed. The log extends Chapter 1’s redundancy test, which asks of a scene whether the reader has already received its shift, to the scale of the book.

The option list (test 2). At four points, after the beginning, at the midpoint, at three-quarters and at the start of the ending, list the moves open to the protagonist towards the goal, including moves they have declined but could still make. Beside each move that disappears between two points, write the page that closes it and whether the closing was by the opposition, by circumstance or by the protagonist’s own choice. A list that does not shrink between two points marks the stretch to revise; a list that shrinks only by the protagonist’s choice marks a middle without opposition; a list with more than one move left at the start of the ending marks an ending that will read as one choice among several. The household lists are in section 7.

The shuffle (test 3). Write each sequence on a card and shuffle the cards. A week later, or with a reader who has not seen the summary, put them back in order using only what is on the cards. Every place where the order cannot be recovered is a place where no cause links the sequences, and where a stations middle has no rule.

The opposition’s calendar (test 5). Write the opposing force’s own timeline through the middle, dated, as if it were the protagonist of its own book, and mark each point at which it surfaces in the novel as a consequence the protagonist feels. Any quarter of the middle with no mark fails the opposition test; any mark that does not change the protagonist’s option list is a reminder rather than a pinch.

6. Further analyses

6.1 Catching Fire: the turn that closes every move

Suzanne Collins’s Catching Fire (2009), the second novel of the Hunger Games trilogy, is in three parts of nine chapters each: “The Spark”, “The Quell” and “The Enemy”. Its first part is a middle of a contained kind, set mostly in the protagonist’s home district and on a tour of the others, in which Katniss Everdeen tries to placate the president she defied at the end of the first novel, chiefly by staging an engagement with Peeta Mellark for the cameras, while unrest spreads through the districts. In chapter 12 the president announces the rules of the Quarter Quell, the anniversary Games: the tributes will be reaped from each district’s existing pool of victors. As her district’s only female victor, Katniss is certain to return to the arena.

The announcement closes every move the first part had opened, placation, flight, private life, at a stroke, and replaces the goal of surviving the regime’s suspicion with the goal of surviving the arena again. It falls in chapter 12 of 27, 12 ÷ 27 or about 44 per cent of the way through the chapters, inside the second part rather than at its start, so that the part’s title has been read for two chapters before its meaning is known. The second half does not break the book’s back because the first half’s question, whether the regime will let her live, is answered by the turn, and answered no.

6.2 The Seven Husbands of Evelyn Hugo: stations held by a frame

Taylor Jenkins Reid’s The Seven Husbands of Evelyn Hugo (2017) is a stations middle in its plainest form. An ageing film star, Evelyn Hugo, gives a little-known journalist, Monique Grant, the exclusive right to write her life, and tells it in seven parts, each named for a husband with an epithet in the title, such as “gullible Mick Riva”. The husbands could be read as an episodic list, and their order is fixed only by chronology. Two engines hold the middle. The first is the frame’s question, why Evelyn has chosen Monique, which is answered only at the end: Monique’s father died in the car crash that also killed one of Evelyn’s husbands, and Evelyn arranged the scene to protect the husband’s memory. The second is a constant hidden behind the list, Evelyn’s long relationship with the actress Celia St. James, which each marriage serves, conceals or damages, so that the husbands are stations on a road whose destination is not any of them. The book shows what a stations middle needs when its stations cannot narrow: a question outside the list that every station brings closer.

6.3 The second stage of Great Expectations

Dickens’s Great Expectations (1860–61) divides its 59 chapters into three named stages of Pip’s expectations, and the division marks a turned middle with unusual explicitness. The first stage (chapters 1 to 19) ends as Pip leaves the forge for London with his expectations; the second (chapters 20 to 39) is a middle of rising social price, in which Pip spends, borrows and grows ashamed of Joe; and the second stage ends with the return of the convict Magwitch, who reveals himself as Pip’s benefactor. The turn falls at 39 ÷ 59, about 66 per cent of the way through, later than the centre, and it is a recognition in Aristotle’s sense, reversing what Pip believed about where his fortune came from, Miss Havisham, and therefore what it was for. The third stage is its consequence. The stage divisions show that Dickens planned the middle as a rise towards a reversal, and that he placed the reversal at two-thirds rather than at half, which a novel can do when its consequences need more room than its build.

6.4 Squid Game: narrowing by turning an alliance into an opposition

The first season of Squid Game (Netflix, 2021), created by Hwang Dong-hyuk, has nine episodes, in which players in debt compete in children’s games whose losers die. Its middle builds alliances among the players as their best move, and the sixth episode, “Gganbu”, turns that move against them. The players are told to form pairs, choose partners they trust, and only then learn that the game will be played between partners, so that one of each pair must die. The episode falls at 6 ÷ 9, two-thirds of the season. It narrows by converting a move into its opposite: the alliance, the move the middle taught the players and the viewer to value, becomes the means of the loss, and the episode’s best scenes are the partners’ choices about what to do with that knowledge. One young woman, Ji-yeong, lets her partner win, and the old man, Il-nam, lets the protagonist cheat him. This is the strongest form of narrowing, the closing of a move by turning it into its opposite, and it is available to any middle that has taught the reader to rely on something.

7. The household middle in full

The Core chapter’s revision case gives the draft middle, the tests, the revised sequence and two passages. This section adds the option lists for each sibling and the canon consequences.

The revised middle runs as follows, positions estimated. At about 25 per cent the drawer chapter ends with the ledger and the looser hand (Chapter 6). At about 30 per cent Dov’s thread, seen from outside, shows the buyer’s demand that the building, shop and flat be emptied by the end of the month (Chapter 14). At about 38 per cent Elena comes back to finish reading and finds the drawer empty; a boy knocks on the glass about the sign, which still reads OPEN; she waves him off, turns the sign to CLOSED, and decides on the way home not to ask Dov who took the ledger. At about 50 per cent, the midpoint, Dov’s messages reach both sisters, asking them to clear the flat by the end of the month without saying why, and Nadia plays hers. At about 58 per cent Dov signs the papers without telling his sisters. At about 66 per cent Nadia, clearing the flat, puts her hand into the pocket of the coat on its hook and finds the good pen (Chapter 7). At about 78 per cent Dov’s chapter, the evening of 9 March (Chapter 13), begins the climax.

The option lists show the narrowing. They list moves towards each sibling’s goal, Elena’s to know why the locks were changed and what the ledger holds, Nadia’s to keep the coat and the flat as they are, Dov’s to sell before another season of losses.

Read across, the table shows each sibling’s list shortened by a different agent: Elena’s chiefly by her own decision, Nadia’s by Dov’s request, Dov’s by his own signature. Read down, it shows the second sag the Core case records: between 66 and 78 per cent Elena’s list does not change and Nadia’s changes only by the passing of days, so that stretch, twelve per cent of the book, is held by the clock alone.

The revision has canon consequences, recorded in the book’s canon file as proposals for the author to confirm. It places the empty drawer, the request and the signing at estimated positions, and settles the order of the request and the signing, which Chapter 14 left open, as request first. It gives Elena a decision not to ask Dov, which keeps her from putting her question to him until the convergence. It turns the sign to CLOSED and gives the prologue’s boy at the glass a second appearance, a boy not identified with the first and waved off in the same way, which pays the two details the prologue left open by rhyme rather than by plot. And it makes Nadia’s unanswered calls from the opening the carrier of the midpoint, which adds nothing to the canon but a voicemail.

8. The limits of the model

The narrowing model describes middles that move towards a single end, and it describes them well. It describes less well three kinds of book. Novels of duration, The Magic Mountain or The Rings of Saturn, prefer the experience of time to the closing of possibilities, and the model will report them as failures of the clock or of the midpoint. Novels of accumulation without convergence, the large social novel whose threads never meet, narrow nothing in particular and gain their force from extent. And comic novels of digression, Tristram Shandy or Three Men in a Boat, delay for the pleasure of the delay. For these the model’s tests still describe what the book is doing, since a reader can see that nothing narrows, but its remedies are wrong, and a writer of such a book should read a failed test as a description, not a verdict.

A second limit concerns the reader. The claim that momentum is the rate at which a reader’s sense of possible endings shrinks is a conjecture, consistent with Brooks and with editorial experience and not measured. Readers differ in how many endings they hold in mind, and genre conventions supply some of them in advance, so that a reader of a crime novel knows from the first page that the question will be answered, and feels narrowing as the reduction of suspects rather than of endings.

9. Sources

Aristotle, Poetics, translated by S. H. Butcher, in Aristotle’s Theory of Poetry and Fine Art (London: Macmillan, 1895); chapters 7, 9 and 11.

Barthes, Roland, S/Z (Paris: Seuil, 1970); translated by Richard Miller (New York: Hill and Wang, 1974).

Bell, James Scott, Write Your Novel From the Middle (Compendium Press, 2014).

Brooks, Peter, Reading for the Plot: Design and Intention in Narrative (New York: Knopf, 1984).

Field, Syd, Screenplay: The Foundations of Screenwriting (New York: Dell, 1979); The Screenwriter’s Workbook (New York: Dell, 1984).

Freytag, Gustav, Die Technik des Dramas (Leipzig, 1863); translated by Elias J. MacEwan as Technique of the Drama (Chicago: Griggs, 1890s).

Mercier, Vivian, review of Waiting for Godot, The Irish Times, 18 February 1956, p. 6.

Reagan, Andrew J., Lewis Mitchell, Dilan Kiley, Christopher M. Danforth and Peter Sheridan Dodds, “The emotional arcs of stories are dominated by six basic shapes”, EPJ Data Science 5 (2016).

Snyder, Blake, Save the Cat! The Last Book on Screenwriting You’ll Ever Need (Studio City: Michael Wiese, 2005).

Wolitzer, Meg, review of The Goldfinch, NPR, 21 October 2013.

Works discussed: Austen, Pride and Prejudice (1813); Beckett, Waiting for Godot (1953; English 1955); Cervantes, Don Quixote (1605, 1615), translated by Edith Grossman (2003); Collins, Suzanne, Catching Fire (2009); Collins, Wilkie, The Moonstone (1868); Defoe, Robinson Crusoe (1719); Dickens, Great Expectations (1860–61); Fielding, Tom Jones (1749); Groff, Fates and Furies (2015); Haig, The Midnight Library (2020); Hemingway, The Old Man and the Sea (1952); Highsmith, The Talented Mr Ripley (1955); Jerome, Three Men in a Boat (1889); Mann, The Magic Mountain (1924), translated by John E. Woods (1995); Melville, Moby-Dick (1851); Reid, The Seven Husbands of Evelyn Hugo (2017); Sebald, The Rings of Saturn (1995), translated by Michael Hulse (1998); Tartt, The Goldfinch (2013); Towles, A Gentleman in Moscow (2016); Woolf, Mrs Dalloway (1925) and To the Lighthouse (1927). Screen: Groundhog Day (Harold Ramis, written with Danny Rubin, 1993); Psycho (Alfred Hitchcock, 1960); Squid Game, season 1, episode 6, “Gganbu” (Netflix, 2021). Towles’s description of his chronology is from the question-and-answer page on his website; the Vanity Fair remark on the reception of The Goldfinch is reported in the novel’s Wikipedia article.

The Middle — Chapter 11, Reference: the full treatment. The reader’s version: The Middle.