Back to the table

72 Beginner Pl Plot Plot & Beats

Plot

The chain of events that makes up the story.

What it is

Plot is the chain of events that makes up a story. Ask ten novelists to define it and you get ten versions of one sentence: plot is what happens, and why. True, and nearly useless. "what happens" fits anything. A character walks to a store. A war starts. A phone rings. All three happen. None of them is plot yet.

A narrower working version exists, and it earns its keep. In the working version, plot is the consequential events only, and every link in the chain passes through someone's choice. An event is something that merely occurs. A plot event is a desire strong enough to make a person act, and an act that costs her options. The gap between those two explains why one busy manuscript grips and an equally busy one comes back with the note "a little slow".

Ansen Dibell gave the most useful version in print. Dibell was a pen name. She lived in Cincinnati, taught writing, edited at Writer's Digest, and wrote Plot for Writer's Digest Books in 1988, in the Elements of Fiction Writing series. The book is listed below. It opens with her definition: plot is made of significant events, and an event is significant when its consequences matter. Showering is not plot. Braiding your hair is not plot, unless the braid is Rapunzel's. The braid matters solely because the prince climbs it. The act of braiding is unchanged. The difference lies in whether later events hang on it. A page later Dibell gives a second definition: plot is what a character does, feels, thinks and says, wherever that makes a difference to what follows. Her one-word summary of plot is a verb.

So the first working definition is the dependency set. Plot is the events that other events rely on. Everything else in the book is texture. Texture is not a fault. A novel needs it. But texture alone will not carry a reader through four hundred pages.

Dibell's second definition does the heavy lifting, because it puts a person inside every link. The pattern never changes: something happens, someone chooses, something happens. Three parts, every time. Drafts lose the middle part. The chain still moves without it, but it moves mechanically, and readers hear the machinery.

Try one example in two versions. A translator is working on a commercial contract. She spots a clause that whoever drafted the source mistranslated, and the error would cost the smaller party a large sum. Version one: the mistake comes to light and there is a scandal. Version one is a sequence of events, and you could swap the translator for a spellchecker without losing anything. Version two: her sole client is the larger party. Her visa hangs on this contract. Nobody else on the premises reads both languages. Her translation reproduces the error unchanged, and she stays silent.

Plot begins at her choice. Everything after it is hers: the money that moves, the business that goes under, the day someone finally reads both columns. The choice cannot be taken back. Each later chapter is a result she owns, and each later chapter is about her whether she is on the page or not. The price of turning version one into version two was roughly one sentence. The information is identical and so is the event. The added element is a person with a stake, choosing between two futures. Drafts that feel lifeless despite heavy incident almost always have this missing: the incidents befall the viewpoint character instead of flowing from her.

The decision is the subject here, the human step between events. If you can name a decision between each pair of events, you have a plot. If you cannot, you have a timeline. Carry one sentence with you through the rest of the page: a plot is a decision chain, and every decision shrinks the character's remaining moves. The mechanics of how one event produces the next, the "therefore" and the "but", coincidence, and endings rescued from outside, belong to cause and effect. Read that page next. None of it is repeated here.

Plot gets confused with four of its neighbors. Story is the bigger entity: someone wants something, something stands in the way, and by the end the person's position relative to the want has changed. Plot is the chain of events that delivers that. A flawless plot with no story reads competent and hollow. A clear story with no plot reads like a stretched synopsis. The story page sets out the difference.

Structure is proportions: where the turns fall within the book. Plot is what the turns contain. Two unrelated novels can share one structure. You can restructure a single novel, reorder its scenes, add a frame, and alter zero plot events. Most supposed structure problems are plot problems. A structure is only a set of positions, and positions filled with something real cannot be wrong.

Premise is the compressed idea: who wants what, against what, and why this book rather than the ones beside it on the shelf. Plot is the premise spent. A premise that never becomes a plot is a very common fault in finished drafts, and it gets its own test further down.

The dramatic question is the single open matter waiting to be settled. Plot is the path to settling it. A strong question with a weak plot produces a reader who knows the stakes and skims anyway. The two fail in different ways, so diagnose them separately.

Seen from a distance, a plot is a sequence of turning points. Each turning point has its own page, so the order is named here once, for bearings only. A beat is one meaningful step. The inciting incident derails ordinary life. The first plot point is the irreversible commitment. Rising action is the long, costly middle. The midpoint changes the game. All is lost empties the plan. The climax is the action that settles the question. The resolution shows what the change cost. If a term is unfamiliar, follow its link. Each one fails in its own ways.

Row 9 of the table holds fourteen named arrangements of those beats: Three-Act Structure, Freytag's Pyramid, the Hero's Journey, the Heroine's Journey, Save the Cat, the Story Circle, Seven-Point Structure, the Snowflake Method, Story Grid, Story Engineering, Kishōtenketsu, the Kübler-Ross Arc, Seven Basic Plots and Romancing the Beat. Almost all of them are beat lists in a chosen order, each with its own vocabulary and its own proportions, and each with a reason for those proportions. Some came out of screenwriting and were adapted for novels later. Some came out of drama criticism. One came out of grief psychology. None of them competes with this page or with each other. A structure labels a plot that exists, or it labels one that does not. It cannot generate one. It can tell you that an event probably sits near the halfway mark; it cannot say what the event is. Where each structure fits and where it strains is on its own page. Comparing them here would duplicate those pages badly.

A subplot is itself a plot: a smaller chain with its own want, its own opposition and its own resolution, running parallel to the main one. The question that matters at this level is where the two chains join. A subplot that never touches the main plot is a second book bound inside the first. Readers sense that it has no bearing on the outcome they care about, and they skim it. A subplot that collides with the main plot at the moment of greatest pressure makes the book feel rich rather than long, because the collision forces a choice. Both chains cannot be served, so one has to be given up. The frequent failure is a subplot inserted to do the main plot's job: the romance solves the murder, or the mentor's side thread delivers the weapon. Readers treat that as the structural cheat it is, and the cause and effect page explains how readers punish it.

You will hear that some writers are plot writers and others are character writers. About nine-tenths of that is nonsense and one tenth is true. Henry James settled the nonsense in 1884. In "The Art of Fiction", in Longman's Magazine in September 1884, he dismissed the then-conventional split between the novel of character and the novel of incident as old-fashioned, and he made the case with two questions: "What is character but the determination of incident? What is incident but the illustration of character?" His point was that you only get to know a person through what they do. Inner life helps, but one paragraph of a character overcoming fear and acting tells you more than a page of her thoughts about courage. A novel of character, he added, made no more sense to him than a picture of character. Decisions reveal character. Decisions make up plot. One substance, two views.

The distinction survives because it describes two genuine working habits. One writer begins with events and derives the actor. Another begins with a person and derives the events. Both routes converge when the destination is clear, and each has a signature failure. Event-first writers force behavior against character for the sake of the mechanism. Character-first writers produce a finely understood person to whom nothing happens. Knowing which way you lean is useful. Treating it as a fixed identity is not.

Now the true tenth. Deliberate plot-lightness can succeed. A superb novel can center on a mind, on a voice, on a place that accumulates page by page, or on a single relationship seen from many sides. Such books still keep a spine; they use a substitute. A confession has a shape: a slow circling toward the thing the narrator will not say. One day inside one head has a shape: a mood building until it snaps. The forward pull persists, and its source is still an outcome with two possible landings. The practical rule: carrying less plot shifts more load onto every other element, and fewer dull pages get tolerated. In a tightly plotted thriller a flat chapter survives, because the reader wants the outcome. A plot-light novel has no such cushion; page quality is the only reason to keep going. Neither option is superior. The choice is a trade. Make it deliberately rather than drifting into it.

The plotter-versus-pantser split is about drafting process, and it says nothing about what kind of book comes out. The most misleading version claims that discovery writers make loose, character-led books and outliners make tight, plot-led ones. In reality, rigid plotting in commercial fiction often arrives without any outline, and bagginess often comes from a plan. The real difference is when you pay. The outliner pays first. Her errors surface in a document, and a wrong turn costs an afternoon. The discovery writer pays later. His errors surface in prose, and a wrong turn costs three chapters, with one bonus: the material arrives alive, where planning might have flattened it. The total cost is about equal. The invoices arrive at different moments.

Most working novelists sit in between, and that position deserves more respect than the word compromise gives it. The middle habits look like this. A fixed ending plus a handful of fixed points, with the route discovered. Outlining three chapters ahead and no further. Writing the synopsis after chapter one, once you know how the book sounds. Planning the plot and discovering the scenes. Or discovering the plot and planning the revision. The industry's nickname for the position is "plantser", which is ugly and accurate.

Each extreme has a hazard. The outliner's is defending the plan against the book: writing a scene on the plan's authority after she has stopped believing in it. Readers notice. An outline is a hypothesis to be tested. The discovery writer's hazard is not getting lost. Everyone gets lost. His hazard is staying unaware of it for eighty pages, which produces a first draft that is really a long and expensive outline. One remedy serves both: scheduled inspections of the spine, never left until the end. The outline page explains how a plan works, and the revision page shows the price of discovering the spine late.

Plot also changes with length. A short story typically holds a single decision and its fallout. The form itself asks for that, and it is no restriction: one choice that cannot be undone, the ground shifting under it, and a closing line that changes what the choice meant from what the character supposed. Most failed short stories are novel openings in disguise, with a decision not yet paid for.

A standalone novel is a closing chain. Whatever important thing opens also shuts, and the readers' word for that is satisfying. Writers overestimate how many of their scenes later scenes rely on. List the scenes that something later depends on and you get a fraction of the total. The remaining scenes are reaction, pressure and texture, arranged around the core. No fault there by itself. The fault appears when every consequential scene sits in the opening quarter or the closing quarter.

A series runs two chains at once. Inside each book, the plot closes within its own covers. That part is mandatory, and public complaints follow any book that seems to hold its plot back for the sequel. Across the set, a larger chain stays open. The big chain needs decisions too, otherwise you have episodes with a logo on them, which is the second-act fault in slow and expensive form. How the two chains relate is on the series page and the series arc page.

Why it matters

Readers talk about plot. When someone recommends a novel, the recommendation never cites prose rhythm. The recommendation is a plot summary: a woman inherits debts from her sister and later ends up working for the man responsible for them. A book that cannot produce a sentence like that gets read, enjoyed, given four stars, and never mentioned again. Word of mouth is plot-driven, so plot's effect on sales is larger than its share of the page count.

Plot sets your drafting speed. You notice this under deadline. Each decision leaves the character a limited few futures, so a plot built from decisions partly writes itself. Say chapter eleven closes on a choice that shuts a door. Chapter twelve has only two or three moves left, and she has to stomach one of them. Compare a chapter that closes on fresh information only. The next chapter could be anything, and nothing drafts slower than anything-goes. So the decision work belongs before drafting, not during it, whatever you think of outlines. The first draft page picks this up.

Plot is the easiest layer to diagnose. Prose is slow to fix and hard to judge. Voice is nearly uncheckable. Plot can be laid out on one page, as the list of the protagonist's decisions, and the point where the plot stops shows at once. Faults become gaps you can point at. Big structural rebuilds, the kind that redo a third of a book, could nearly always have been small with a mid-draft spine check.

Plot is where character becomes visible. People are known by their costly actions. A plot is a set of occasions engineered to make the protagonist reveal herself. A plotless book needs another route to knowability. That route exists, and it is much harder.

Five faults account for most of a developmental editor's margin notes. A developmental editor is one who judges the big picture of a story rather than its sentences. Each of the five has a test you can run this week, privately. If your second act sags, James Scott Bell's Plot & Structure, listed below, is the standard practical handbook, and its chapter six on middles and chapter fourteen on frequent plot faults and their fixes both work as exercises.

The episodic middle. The chapters could run in nearly any order. The prose is fine and each chapter is enjoyable on its own, yet the book reads long. Ambitious first novels have this structural fault more often than any other, and it sits almost always in the second and third quarters. The test is the swap test. Physically reverse two middle chapters and read the join. If nothing breaks, no reference to an event that has not yet happened, no wound nobody has yet suffered, no fact used before its discovery, you have episodes rather than plot. Repeat with three more pairs. Mostly clean swaps confirm what the reader felt: nothing accumulates. The repair: every episode must subtract something, not only add information. It can subtract an option, an ally, a resource, a hiding place or a version of the plan. An outcome that narrows nothing is no link yet.

The carried protagonist. She is present everywhere and responsible for nothing. She is invited, summoned, rescued, recruited, told, warned and driven. Readers call her flat, unlikable, hard to root for, and writers respond by revising her personality. The fault is structural. The test is a decision count. Take the first hundred pages and list each decision that meets two criteria: she paid a price for it, and she had a plausible alternative she might have chosen. Leave out preferences and reactions. Fewer than four entries means she is a bystander to her own plot. The harder version is a substitution check: would a different character, making the same choice, produce an identical outcome? If so, you have an errand rather than a choice. The repair is typically one early line. Someone offers her an exit and she declines. The agency page gives the full treatment, but the point belongs here too: passivity in a protagonist has a structural cause and character-shaped symptoms, not the other way round.

The information ending. The problem ceases because of an explanation. A letter arrives, a witness speaks, a will is read aloud, someone confesses, an expert names the compound, and the protagonist is present only as the recipient. The test is a sentence test. Go to the last chapter and find the line at which the main problem is over. Note who speaks it and what the speaker does. If the speaker is not the protagonist, or the sentence is a document read aloud, or it opens with "what you don't know is", the verdict is delivery. A true climax has the protagonist act, pay, and use something planted earlier. The repair: move the information earlier and deliver it in a form that demands action, so the last chapter becomes the act. Information may serve as the turn, never as the resolution.

The pressure reset. Tense chapters, slack book. Trouble gets handled, and the next chapter opens at the old position: same resources, same options, same relationships, only more fatigue. This one is harder to see than the episodic middle, because the causal links exist. The accumulation does not. The test is a standing-position log, one line per middle chapter answering what can she no longer do that she could do at the start of this chapter? Three chapters in a row with identical answers, or no answer, confirm a reset. Add a reverse column: what can she now do that she could not before? A healthy middle oscillates, with net movement downward until the final stretch. The repair: stop fully solving chapter problems. A full solution returns her to the start. A solution at a price advances the book. She escapes, but her only possible character witness now refuses.

The premise repeated. Of the five, this one is the most disheartening to find, and the commonest in books with a strong idea. The concept is strong, and each chapter shows it again. A hospital where every patient's death is announced a week in advance, followed by forty chapters of staff struggling with that. A marriage collapsing, shown in thirty scenes the same way. Nothing develops, because the engine is a standing situation rather than a chain. The test is a heading test. Write a one-line summary of each middle chapter and group the summaries under headings. Fifteen summaries under a single heading that restates the premise means illustration, not development. The second symptom: each chapter could be published alone, and the order is irrelevant. The repair: a character inside the situation wants out, acts, and makes things worse. That switch, from resident of the situation to its opponent, is the switch from premise to plot. The premise page gives the other side of it, and the stakes page explains why "worse" has to be specific.

Real-world examples

Atonement (2001): one child's decision and the four hundred pages that hang on it

Ian McEwan's Atonement, published by Jonathan Cape in 2001 and shortlisted for the Booker Prize that year, is the clearest modern demonstration of one choice generating a whole plot. The smallness of the choice is what makes it worth study. In the summer of 1935, Briony Tallis is thirteen. She has read a note Robbie Turner wrote to her sister Cecilia and trusted Briony to carry. She has witnessed two encounters between them that she does not understand. That night her cousin Lola is assaulted in the dark. Briony did not see the attacker. She claims she did, tells the police, and holds to it. Robbie goes to prison on her word.

Nothing was done to her. There is no compulsion and no threat, and at every moment the available alternative is "I couldn't see". The root cause of everything that follows is a child's decision to treat suspicion as certainty. Four hundred pages hang on it: prison, release into the army, Dunkirk, the nursing wards, and Briony's sixty years of writing. One page of the novel contains the entire difference between an event and a plot event.

Non-decisions chain too. Across the following years Briony's certainty about Paul Marshall, the real attacker, grows. She never retracts. Each year of silence repeats the same choice, so the pressure is built from a renewed decision rather than a fresh one. The same pattern transfers to any protagonist who is free to end her situation and keeps not ending it.

Skip the next paragraph if you intend to read the novel, because the final twenty pages recast everything before them. In the coda, set in 1999, Briony is seventy-seven and is revealed as the author of the novel you have just read. The reunion between Cecilia and Robbie never took place. Robbie died of septicemia at Dunkirk. Cecilia was killed months afterward in the bombing of Balham Underground station. Read as a plot point rather than a trick, the coda is the same character making a second decision with the identical coin as the first. Her first alteration of the record served herself. The second serves kindness. The final turn of the book has the protagonist repeat the act the whole book concerns, which shows plot and character to be one material.

Adventures of Huckleberry Finn (1884): an episodic middle that works and an ending handed to a will

Mark Twain's Adventures of Huckleberry Finn was first published on December 10, 1884 in the United Kingdom and Canada, and on February 18, 1885 in the United States. The polite label for its form is picaresque: loosely joined adventures in a string. A novelist should study it for two reasons. Its episodic middle succeeds, and its final eleven chapters are the most disputed example of a resolution delivered rather than acted.

Run the swap test on the middle. Huck and Jim go downriver. They are separated in fog. There are the Grangerfords and their feud. Two con men board the raft, one claiming a dukedom and the other claiming to be the lost dauphin of France, and there follow the scams, a camp meeting, a mangled Shakespeare performance and the impersonation of the Wilks brothers. Reorder several of those and almost nothing breaks. Some could be removed without harming the chapters on either side. The picaresque bargain works like that, and here the payoff is huge: the river is the vehicle that lets a boy see a whole society in cross-section, and that is the novel's purpose. What glues it is the single non-episodic element. The raft travels one direction only. As it goes downriver, Jim moves further from freedom, and the bond between Huck and Jim deepens even when the incidents do not.

The payoff arrives in chapter 31. Huck writes the letter that would return Jim to slavery, reflects on what the two of them have been through, and tears it up. That "all right, then, I'll go to hell" moment is the true climax, and it ranks among the great decisions in American fiction. A choice, with a price, his whole belief about his own damnation, and irreversible.

Then the final eleven chapters hand the resolution to information. Tom Sawyer arrives and stages a needless, elaborate "escape" for Jim. The reveal: Miss Watson, who died two months before, had already freed Jim in her will, and Tom knew. In retrospect Huck's choice weighs nothing. A document had already achieved the thing he accepted damnation for. In Green Hills of Africa, in 1935, Ernest Hemingway advised readers to stop at the point where Jim is stolen away from the boys, since that is the real end and the rest is cheating. T.S. Eliot and others have defended the closing chapters as suited to the book's shape. The argument has run since 1935 and there is no need to settle it here. The lesson is in Hemingway's reaction. He was a working novelist judging a working novel, and his best available reading is that a resolution by revelation cancels the decision that built the plot. When a reader reports your ending as a letdown and cannot say why, check this first.

Architects and gardeners: Deaver, Martin and Child on how a plot gets made

The claim that outlining and discovery produce different kinds of book can be checked against first-hand testimony from both extremes.

Jeffery Deaver is the planning end. In a Q&A with Robert Lee Brewer, Senior Editor of Writer's Digest, published December 19, 2022, Deaver said that more than half his working time on a book goes to the outline, before any prose. His outlines run a hundred pages, sometimes a hundred and fifty. His reason: fast, plot-driven books require a structure that holds, and wrestling the plot before drafting exposes what fails while it is cheap to fix. His novels run several subplots at once inside a story span of a few days. Clockwork of that kind is nearly impossible to discover by instinct.

George R. R. Martin is the gardening end, and he popularized the labels. In "A Winter Garden", posted on his site Not a Blog on July 8, 2022 and listed below, he calls himself a gardener and is candid about the cost. His stories alter as he writes them. He knows the destinations and the major set pieces, and not the route. The ground moves under him, and he describes repeatedly tearing everything apart and rewriting. Some doors go nowhere. Some open onto marvels.

Lee Child is the extreme gardening case. Andy Martin of the University of Cambridge sat behind Child while he wrote a Jack Reacher novel and reported it in The Conversation on November 27, 2015, also listed below, and offered as the shortest answer to the plotter-versus-pantser debate. Child starts every book on September 1, the day he bought pencil and paper for book one. For the observed novel his starting material was a title thought up the night before plus one picture in his head of men burying somebody. No outline. No notes. No plot. He completed it in one draft.

Put the cases together and they exceed anecdote. Deaver and Child write in the identical genre at identical length for a largely shared readership, and their finished plots are alike: tight, driving, engineered. Deaver spends months in a document and types the prose in weeks. Child's engineering emerges in the prose, with no revision. The process is opposite and the product is the same. Martin, a gardener, produces fantasy plots as tightly interlocking as any in the modern genre, and pays for them in restructuring. So the plotter-versus-pantser dispute is not about the shape of the book. Tolkien supplied the top single sentence ever written about discovery writing, in a letter to W. H. Auden on June 7, 1955, about writing The Lord of the Rings. Strider's arrival at the inn came to him as a shock, and he knew exactly as much about the stranger's identity as Frodo did, which was nothing. A plot discovered is a plot nonetheless. Its author merely read it first.

Counting the plots: thirty-six, twenty, seven, four, and two and a half thousand

Sooner or later you will meet the claim that there are only so many plots. The count varies wildly by author, and that variation is the exercise's main lesson. The credits are often wrong too, so here they are set out carefully.

Georges Polti, a French writer, published Les Trente-six Situations dramatiques in 1895. Lucille Ray's English translation, The Thirty-Six Dramatic Situations, came out in 1916 and is out of copyright, so the full text is listed below, and you can decide for yourself whether such a catalog helps a working writer. Polti saw himself completing an older claim. Goethe had reported Carlo Gozzi (1720 to 1806) as holding that only thirty-six tragic situations exist, and Schiller had searched hard for more and could not reach even thirty-six. Polti worked through classical Greek drama and much French theater to fill the list. The thirty-six are situations, not plots: arrangements of pressure such as supplication, vengeance, fatal imprudence and rivalry between kin. A novel usually chains several in sequence. The list shows where it came from, the French stage of the 1800s and classical drama.

Ronald B. Tobias's 20 Master Plots and How to Build Them, from Writer's Digest Books in 1993 and revised in 2012, is the most practically useful catalog, and the count has nothing to do with why. The patterns are the expected ones (quest, pursuit, rescue, escape, revenge, rivalry, temptation, transformation, forbidden love) plus sharper ones such as the riddle, wretched excess, and the paired ascension and descension. Tobias treats each pattern as a set of obligations owed to the reader of that pattern. Choosing "pursuit" commits you to what pursuit readers demand. The obligations turn the catalog into a checklist, and a checklist is something you can use.

Christopher Booker, a British journalist, spent thirty-four years on The Seven Basic Plots: Why We Tell Stories, published by Continuum in 2004. His seven are Overcoming the Monster, Rags to Riches, The Quest, Voyage and Return, Comedy, Tragedy and Rebirth, argued along broadly Jungian lines. His case and the objections to it live on the Seven Basic Plots page and are not repeated here.

Jorge Luis Borges got down to four. His short essay "Los cuatro ciclos", in his 1972 collection El oro de los tigres, names four stories: a city under siege, with Troy as the example; the return home, Odysseus; the quest, with Jason and Ahab both under it; and the sacrifice of a god, the Crucifixion. He was not proposing a writing method, and the popular versions credited to him sometimes list four different things.

The scholarly answer is deflating: around two and a half thousand. The Aarne-Thompson-Uther index began as Antti Aarne's Verzeichnis der Märchentypen in 1910. Stith Thompson translated and enlarged it in 1928 and revised it again in 1961. Hans-Jörg Uther expanded it as The Types of International Folktales in 2004, in three volumes. Folklorists use it as a finding aid and claim no theory for it. It has been criticized for European bias and for prudishness about its material, and Vladimir Propp objected that it files tales by motif and never asks what a motif does inside the tale. Propp's own work is covered on the archetype page.

The nearest thing to a data-driven answer is six, and it concerns emotional arcs rather than plots. A 2016 study in EPJ Data Science by Andrew Reagan, Lewis Mitchell, Dilan Kiley, Christopher Danforth and Peter Sheridan Dodds identified six repeated shapes in the emotional trajectory of stories. The shapes recur, but what they measure is feeling over time, not plot. The story page discusses them properly.

So the running total is four, six, seven, twenty, thirty-six and two and a half thousand. Two things are true together. The catalogs are not useless. Tobias's obligations diagnose, and knowing the revenge plot tells you the reader expects the avenger's moment of chance plus the choice to use it. And no plot has ever once been generated from a catalog. Catalogs face backward. They summarize what stories became, and each slot holds thousands of titles. Selecting "the quest" from the twenty gets you as close to a novel as a swatch on a paint chart gets you to a painting. The correct use is confirming the promise you have made. The wrong use is generating a plot.

Gilead (2004): a near-plotless novel with a spine

Fairness demands a celebrated near-plotless novel, so take Marilynne Robinson's Gilead, published by Farrar, Straus and Giroux in 2004, winner of the Pulitzer Prize for Fiction in 2005 and of the National Book Critics Circle Award. The form is one long letter, written in 1956 by the Reverend John Ames, seventy-six, dying of a heart condition, to his seven-year-old son, who will barely remember him. He writes about his abolitionist-preacher grandfather, his father, his short first marriage, his late second marriage, and light on a wall. As plot summary that is nearly nothing. There is no conventional inciting incident, no midpoint reversal, no confrontation at the climax.

The book is not shapeless, and its shape can be borrowed. One thing looms from page one: Ames's coming death, with the letter as his only legacy. And there is one subject he avoids: Jack Boughton, full name John Ames Boughton, the son of Ames's closest friend, named after Ames, with Ames as his godfather, a man Ames has never trusted, now returned to town. The avoidance is the engine. Ames approaches the subject, retreats into memory, and approaches again. The pressure comes from a confession postponed while time runs out. The reader's question is whether Ames will manage generosity toward Jack before he dies, and that meets the dramatic question standard: a yes-or-no the ending can deliver. The ending delivers it, with the blessing.

So Gilead has a spine. Outside events are replaced by successive approaches to the unbearable, and the deadline is the narrator's failing body. The lesson for quiet fiction: no exemption from the pull. The pull has to come from somewhere else, and the margin for error is smaller, not larger. The extreme case is Nicholson Baker's The Mezzanine, from 1988, roughly 135 pages, told largely in footnotes, whose surface action is an escalator ride back from lunch. There attention itself is the event. Very few writers can hold attention that way, and there is no disgrace in not being one of them. The harm comes from assuming a draft with no pull belongs to that tradition when the chain has simply been lost.

An everyday scenario: a ten-line list for a 108,000-word draft

Picture a writer with a finished fantasy draft of 108,000 words. Two readers have read it. Both are polite, both say it "loses its way in the middle", and both liked the last fifty pages. Instead of rereading, she opens a blank document and lists one decision per line, verb first, nothing else on the line. Events and discoveries are left out. Only costly choices count. The result covers thirty-eight chapters in ten lines.

Her protagonist, Linnet, a traveling woman, leaves home rather than take the apprenticeship. She lies to the toll clerk. She gives away the letter. Then nine chapters pass with nothing for the list. She agrees to guide the delegation. Seven more chapters pass with nothing. Then, in the last quarter, six decisions in a row, which explains why both readers approved of the ending.

The diagnosis took forty minutes. The first act works, because she chooses. The final act works, because she chooses. Chapters nine through twenty-five are a traveling woman to whom things occur. Nobody could name the problem, because every chapter is fine on its own: weather, danger, a good bridge scene.

The fix is smaller than rewriting the sixteen decision-free chapters. For each existing event in the stretch, she asks who decided it. Chapter twelve, as drafted: bandits take the horses. No decision. Revised: Linnet spots the bandits, and her choice is to send the boy down the wrong road with the horses and attempt no defense. The boy is saved, and the fastest route to the pass is lost. Same chapter, same bandits, same outcome on the page, and now it is a link. Chapter nineteen, as drafted: a stranger reports that Osgood, her rival, reached the city first. No decision. Revised: she chooses whether to tell the delegation, and withholds it, for defensible reasons, so a new price lands in chapter twenty-three. She applies the conversion six times across the middle. The book gets no longer. And the swap test now fails, because each chapter removes something the next chapter needs her to lack.

Try it yourself

This takes about one evening for a novel. It works on a finished draft, a stalled draft, or an outline you do not trust. Nothing gets rewritten yet, on purpose: the pass finds out what the plot is before anything changes.

  1. List the decisions, verb first. Open a single document and write one line per decision, in chronological order, verb first: "Refuses the deal", "Tells her brother the truth", "Goes back for the dog". A decision qualifies when an alternative existed and the alternative would have led somewhere else. Discoveries, arrivals, attacks and revelations do not qualify, however dramatic. Read for choices, not prose quality. If you want a fuller card for each entry, use this one and fill in the later lines as you go through the next three steps.

Decision N: __________ (verb first) Option it closed: __________ Made necessary by: __________ Chapter: __ of __ If deleted, does the book still stand? yes / no

  1. Put both totals at the top. Record the decision count and the chapter count at the head of the document. There is no correct ratio, but under one meaningful decision per three chapters means a passenger protagonist, and a middle with zero has located the problem for you.
  2. Write what each decision eliminated. Next to each decision, name the option it closed: a helper now lost, a route now shut, a fact she can no longer pretend ignorance of. An empty column marks a mere preference, which is not plot. Strike it and see what remains.
  3. Find the real point of no return. Work from the top down and ask of each decision whether the book survives its deletion. The first "no" is the real point of no return, whatever page it sits on. If it sits at page 180 in a 400-page book, you have a hundred surplus pages of setup. Compare the result with the first plot point.
  4. Locate the final decision. Its correct place is the climax, and it should be the hardest. If the last real decision comes three chapters before the end, the resolution has overgrown. If the climax contains no decision, go back to the information-ending test, because the climax is almost certainly a delivery.
  5. Mark the gaps. Four or more consecutive chapters without a decision are the stretch readers call slow. Write down those chapter numbers. Those chapters are your work list. Expect one or two such stretches; ten would be unusual.
  6. Run the swap test inside each gap. Reverse two chapters and read the join. If it holds, you have episodes. This takes five minutes and gives you the most persuasive evidence you can give yourself, because a chapter indifferent to its position is hard to argue with.
  7. Convert events to decisions. For each chapter on the work list, find the existing event and ask who decided it. If nobody did, transfer the decision to her and attach a cost. The usual scale is a page added and a page changed, no new chapter. Aim for one decision per gap chapter. After three or four conversions the entire stretch reads differently.
  8. Read the list aloud, first to last. For each adjacent pair, say why the second became necessary. If you have no answer, the two are unconnected and one of them is floating. Here the plot pass becomes cause and effect work, and the finer instrument lives on that page.
  9. Keep the list at your desk. One page, and it is the book's actual plot. It becomes the reference for every drifting chapter through the rest of the drafting. If you plan, next time outline this list rather than scenes: before chapter one, write the ten decisions the protagonist will face, in order, each with its cost. A full list means you can draft forward without an outline. Stalling at four means you have a premise and not yet a plot, and finding that out costs an afternoon at the outline stage instead of eighty thousand words.

Common mistakes and misunderstandings

Mistaking volume of incident for plot. The busiest manuscripts are often the deadest. Chases, betrayals, explosions and revelations are events only. They become plot once the next chapter is forced to differ because of them. A draft labeled slow may prove, on inspection, frantic.

Treating plotting as a substitute for having ideas. Real novelists hold this view, and it rests on confusing plot with formula. Structure can be formulaic. Consequence cannot, because consequence is just what any book has when things matter. Discovery writers, however committed, end up with plots too. Theirs arrive in prose and get recognized afterward, which differs from having none.

Using structure in place of plot. The middle is broken, so a beat sheet gets applied, all fifteen beats land in the correct positions, and the middle is still broken. The sheet can locate a reversal. It cannot say what reverses. When a structure is fitted and the book is still dead, stop tuning proportions and return to the decisions.

Treating a twist as the plot. A reversal is a single link, never the whole chain. A book built on one late reveal gives the reader nothing to do beforehand, and then the reveal lands weakly, since nothing built toward it. See plot twist and foreshadowing.

Adding subplots to repair a thin main plot. Three extra subplots cannot repair a spine. They conceal the fault for fifty more pages, and then the ending gets harder, with four things needing to land. Repair the main plot inside the main plot. Afterward, if the book needs weight, add a subplot that collides with it.

"My genre doesn't really need plot." Every genre needs consequence. Only the currency differs. In romance the currency is what the pair do to each other irreversibly. In the quiet kind of literary novel the whole plot may be one withheld conversation that lasts two hundred pages. In a cozy mystery the plot is highly visible and highly conventional, with readers counting the beats. The claim usually describes a draft in which events have ceased.

Prescribing more outlining as the cure. Sometimes the outline is not the problem. A detailed plan whose draft dies at chapter twelve gains nothing from more detail. The fault is the type of thing planned: scenes and settings, when it should have been choices and costs. Redo the plan as ten decisions for the same book and check whether it survives.

Keeping the plot to yourself. You know chapter fourteen matters because of chapter thirty-one. The reader does not. No reader extends credit across four hundred pages against consequence promised later. Each chapter has to pay immediately.

Using a taxonomy as a generator. Picking "revenge" from a list of twenty is good for checking the promise you are making. For finding the book's events it is useless. The catalog comes after the fiction, not before.

Postponing the spine to revision. Mid-draft, the list costs one evening. At the end it costs two months, because by then the chapters exist, the prose is good, and you are attached. Run the list at the fifty-thousand-word mark, while what it finds is still easy to act on. The revision page shows the costly version.

Mistaking a synopsis for the spine. A synopsis narrates events in sequence. Its job is sales and planning. Smooth prose can carry an episodic middle through a synopsis unnoticed, which is why a broken book can still yield a fine synopsis. The spine lists choices with their prices, and it cannot conceal a gap.

Confusing plot with pacing. The complaints sound similar and the fixes are opposite. When the reader says nothing seemed to matter, the problem is plot. When the reader says the stakes were clear but she skimmed anyway, the problem is pacing. Shorter chapters do not fix plot, and a rebuilt plot does not fix a flabby scene.

Related elements