Back to the table

57 Structure Ta Three-Act Structure Story Structures

Three-Act Structure

Setup, confrontation, and resolution.

What it is

Three-act structure is the simplest useful way to split a story into parts: a beginning, a long middle and an end. The beginning commits someone to something they want. The middle keeps it from them. The end settles the matter.

Think of the whole story as one continuous movement with two changes of situation inside it. Act one sets up a person, a world and a want. It ends when that person's old life is closed to them. Act two is the chase, with the thing they want always out of reach. It ends when the only option left is the hardest one. Act three is taking that hardest option, followed by a short stretch where things settle down.

Everything else you'll hear about three acts came later. The page counts, the percentages and the named beats are all attempts to pin down where those two changes fall.

Two different ideas share the name, and you need to keep them apart. As a description of how stories feel, three acts is very old. As a prescription, with the act breaks placed at fixed page numbers, it was invented by twentieth-century screenwriting, and Syd Field's book Screenplay spread it in 1979. Mixing the two causes real trouble. Most of the version you'll find online is the prescription borrowing the prestige of the description. Held as the older, looser idea, three acts gives you the fastest and easiest structural test there is. It shows you whether your opening is overloaded, whether your middle has a spine, and whether your ending was earned.

The beats writers talk about sit inside the acts like this. The Inciting Incident happens inside act one. The First Plot Point is the exit from act one. The Midpoint sits in the middle of act two. All Is Lost comes just before you enter act three. Act three holds the Climax and then the Resolution. Each of those beats has its own page. What you need here is the container that holds them: where it came from and how to use it.

The usual proportions are a quarter, a half and a quarter. Act two is as long as acts one and three put together. The size of act two is the weightiest fact about the model, and it causes most of the model's trouble.

The intuition and the method have separate histories. Blending them is the most common error in craft writing about structure, so the history is worth a few minutes of your time.

Start with the intuition. Aristotle did not describe a three-act structure. In the Poetics, from the fourth century BC, he says a whole has a beginning, a middle and an end, and he defines each one by cause. A beginning needs nothing before it but has something after it. An end follows from something and has nothing after it. A middle has something on both sides. He's making an argument about causes and completeness. He doesn't divide a play into units that get performed, he gives no positions, and the word "act" never appears. You can read the passage, 1450b, free at the Perseus Digital Library.

When Aristotle does divide up tragedy, at 1452b, he lists four named parts: prologue, episode, exode and choral song. There are no acts. He also splits a plot into two halves, the complication and the unraveling. The closest he comes to a modern structural claim is his view that reversal and recognition should grow out of the plot's own logic instead of being imposed from outside. As the Climax page shows, he had no word for the modern climax at all, and the up-and-down diagram people credit to him isn't his.

Act divisions are Roman, and the number that settled was five. Horace, in the Ars Poetica of about 19 BC, says at lines 189 to 190 that a play which hopes to be revived should run to exactly five acts, no more and no fewer. No older rule about the number of acts survives, so five has the oldest pedigree of any count.

Gustav Freytag also used five parts in Die Technik des Dramas (1863): introduction, rise, climax, return or fall, and catastrophe, with three "moments" sitting between them. He was studying classical five-act tragedy. His peak comes in the middle of the play, and his last part is a catastrophe instead of a tidy ending. The pyramid you may have learned in school isn't quite his, and it certainly isn't three acts. The Resolution page shows the point about his fifth part from his own text, and Freytag's Pyramid covers his model itself.

Three-act plays came long before three-act theory. Ibsen's A Doll's House, published in Norwegian in December 1879, is a play in three acts, and three-act staging was ordinary in the European theater of the 1870s. But act breaks on a playbill only tell you where the intermissions go. They say nothing about where the turning points fall or what each act has to do, and those are the things a writer needs from a model.

The prescription belongs to Syd Field, and it dates from 1979, when Delacorte Press published Screenplay: The Foundations of Screenwriting. Delta brought out a revised US edition in 2005. The book grew out of a practical problem. In the early 1970s Field ran the story department at Cinemobile Systems. By his own account he read about two thousand screenplays there and passed about forty. He wasn't asking "what is a story." His question was closer to "why do almost all of these two thousand scripts fail in the same way, and can I teach a novice to avoid it." After Field's death in 2013, Alex Pappademas wrote about him in Grantland and called him the first person to publish a mass-market screenwriting guide that set out three acts hinged by plot points.

Field's one-page diagram, the Paradigm, is still on his site, and it's the primary source for what he claimed. It labels the three acts set-up, confrontation and resolution, and it cuts Act Two into two halves, each with its own dramatic context. In a 120-page script, Plot Point I falls roughly on pages 20 to 30, Plot Point II roughly on pages 80 to 90, and the midpoint near page 60. Reading the worksheet teaches you two things a summary won't. First, the table's own summary phrase, "setup, confrontation, and resolution," is Field's vocabulary from 1979, not Aristotle's. Second, Field gave ranges, not single points. His window for Plot Point I is ten pages wide, which is more than 8 percent of a 120-page script. The exact "page 25" and "page 27" that craft blogs credit to him appear nowhere on his worksheet. The First Plot Point page goes further into this.

So what about the claim that three acts is "the oldest and most widely used story shape"? The "widely used" half holds up. Anglophone commercial fiction and film use the model everywhere, and nobody disputes that. The "oldest" half holds only for the intuition. The idea that a finished thing has a beginning, a middle and an end goes back at least to the Poetics, and perhaps as far back as stories have been told. The method, with three acts and named breaks at set positions used to plan and diagnose a script, goes back only to 1979. If someone tells you Aristotle invented the three-act structure, they're making a false appeal to authority, and a false appeal makes a poor foundation for a book.

The model suits a story with a single protagonist, one main want, a mostly chronological telling, and an ending that answers the question the opening asked. Most commercial fiction fits that description: thrillers, crime, romance, most fantasy and science fiction, and most upmarket book-club novels. Where a genre's contract with its readers is strong, the model almost disappears, because it's simply the shape those readers expect. Reader Promise covers that contract.

The model strains in five places. The first is the series. A book that is complete in itself but also opens a series has to answer its own question and start a bigger one. The container handles the first job only, and Series Arc handles the rest.

The second is the ensemble or multi-strand novel. Give a book four protagonists and you have four wants and four act structures. The breaks for the whole book fall where the strands collide. The model works for each strand on its own. A single three-act spine for the whole book is a fiction you keep up in your notes, with nothing on the page to support it.

The third is the mystery. A detective novel's middle piles up clues instead of building confrontation. The Christie example below shows how, and Mystery has the full picture.

The fourth is any shape that isn't driven by conflict at all. That strain is the serious objection, and it comes up under why the model matters.

The fifth is the short form. A short story is short because it has one movement and one turn. Force three acts onto four thousand words and you get a tiny novel instead.

Why it matters

The model is vague, and the vagueness is why it lasted. The usual objection says three acts is too vague to be useful. The vagueness is real, but the conclusion is wrong. The looseness explains why the model has held up since 1979, while templates tied to page counts have aged. Use it to diagnose a draft. Nobody ever wrote a good book just by deciding to use three acts.

Three questions catch most structural problems. The model gives you three questions you can ask of a synopsis, a finished draft or a stuck manuscript, in any genre and at any length. First, is act one doing too much? If the opening quarter is still introducing people and explaining the world, nothing has committed your protagonist yet, and you're asking the reader for patience when you should be giving them something to be interested in. Second, does act two have a spine? The middle is half the book. If you can't name your protagonist's through-line and the force pushing against it, the middle will sag, and readers tend to quit around the third chapter in which nothing happens. Third, was act three paid for? A climax won with a skill, an ally or a fact that showed up in the last twenty pages reads as a rescue instead of a result. Five minutes with a chapter list will catch a lot of faults. For a model you can state in one sentence, that payoff is large.

Find the act breaks by what your character can no longer do. Never find them by counting off 25 percent of your word count. Ask the act-break question of every scene that might be a break. For the first break, ask: after this scene, is the protagonist's old situation still open to them? The test is availability. The question isn't "do they want it" and it isn't "is it wise." If they could still go home, go back to the job, leave the letter sealed or take back what they said, act one is still going, whatever the page number. For the second break, ask: is there anything left to try except the hardest thing? If easier options remain, you're in act two and the book is still building. Once every easier route is shut, act three has begun, whether you meant it to or not.

Your breaks probably aren't where you think. Run the question on your own draft and the breaks usually land somewhere other than where you assumed. Two patterns turn up often. In the first, the first break comes late, because the scene you marked as the turn is one the protagonist could still walk away from. In the second, you find three candidate breaks in a row heading into act three. They're the same escalation written three times, a Pacing fault that the three-act frame is very good at exposing.

Act two isn't a real act, and that explains most of this table. The model leaves one big empty space in the middle, and the heavy lifting is filling it. Saying so takes nothing away from the model. Instead it explains why most of the tools on this table exist. A beat sheet runs to fifteen entries because three can't cover a middle a hundred and twenty pages long. Rising Action is the overall shape of the middle, with the opposition growing while the protagonist's options shrink. The Try-Fail Cycle is the loop of attempt, result and changed position, and the results have to differ, or you end up with five copies of one scene. The Midpoint is the pivot that keeps the back half from repeating the front half. Scene and Sequel is the beat of action and reaction at chapter scale, and it keeps a long middle easy to read. A Subplot adds other strands for variety without throwing the main line off course. Stakes and Conflict tell you what each failure costs and who or what is standing in the way. Cause and Effect is the chain that makes the middle read as consequence instead of one thing after another. The Dramatic Question is the one open matter that holds the reader all the way through. When your act two sags, three acts can tell you so, but the fix has to come from those elements. The model stops there, and saying so plainly beats pretending three acts is a complete method on its own.

Five serious objections deserve your attention. Be suspicious of anyone who presents three-act structure as settled. Five objections deserve a real hearing, and knowing them will make you better at using the model.

The first objection says a description got sold as a prescription. Field arrived at his paradigm by reading scripts and noticing what they shared. In the book he called it a model, not a formula, and he drew that line on purpose. Later, his observation hardened into rules with a precision he never claimed. The strongest form of this objection comes from Kristin Thompson, in Storytelling in the New Hollywood: Understanding Classical Narrative Technique (Harvard University Press, 1999). She studied a large sample of Hollywood films and concluded that the three-act model misdescribes how they're built. In its place she offered four sections of about equal length: setup, complicating action, development, and climax plus epilogue. Her finding goes well beyond a quibble. Once a scholar shows that the model fails as a description, it loses its main claim to authority as a prescription.

The second objection says act two isn't an act. Thompson's four parts come from halving Field's confrontation. Cut it in two and you get four sections of similar length, with a real break marked in the middle. John Yorke, a former BBC drama executive, lands close to her from inside the industry, in Into the Woods: A Five-Act Journey Into Story (2013). He frames his five acts as a refinement and doesn't mean them as a rebuttal. On his site he describes five acts as three acts refined, giving you more control through five divisions instead of three. Both make the same point. A middle that fills half the book with no joints inside it isn't really a unit of structure, and it leaves you to improvise.

The third objection says the numbers are falsely precise. Field offered ranges. Blake Snyder, in Save the Cat! (2005), gave a single page number for each beat in a 110-page script. The Save the Cat brand's present guidance recasts several of those beats as percentages that differ from the book, and the First Plot Point page records a drift of nearly three percentage points on one beat. A screenplay page is a convention worth roughly one minute of screen time, so a feature runs about two hours. A novel page measures nothing. Turning "page 25 of 110" into "22.7 percent of your manuscript" gives you a figure that looks exact and means nothing.

The fourth objection says the shape belongs to one culture. This objection deserves the most attention and gets brushed aside most often. Matthew Salesses makes it in Craft in the Real World: Rethinking Fiction Writing and Workshopping (Catapult, 2021). He argues that Western craft teaching treats one shape as neutral: a chain of causes, driven by the protagonist's desire, in three acts. In his account, that shape is one tradition's audience expectation, handed down from Aristotle and Freytag and presented as the way stories work. He points to four-part East Asian structures, where the third movement is a surprise that changes how you see everything, instead of a confrontation. He adds that writers from other traditions get told their work has no structure, when it simply has a different one. The counterpoint is that for Anglophone commercial readers, the model matches what they've been trained to expect. The fit is a market fact, and a market fact is a separate thing from the claim that this is how stories work. The Beat Sheet page holds his fuller argument, and Kishōtenketsu is this table's counter-case.

The fifth objection says the model was built for a different form. Every detail of the prescription came from feature screenplays of about two hours. The act lengths, the plot-point windows, the midpoint and the unit of measure all started there. A novel has no fixed length, no duty to keep a stranger in a seat and no rule of one page per minute. Tim Lott made this point in an extract posted on his site on January 10, 2023. He wrote that the three-act principle works for drama and screenplays, but the novel is baggier and wants the idea held more loosely. He keeps it anyway. His Yes! No! But Wait…! (Swift Press, 2023) builds plot, character, theme and causality on a three-act base and closes with a Structure Grid. His book is listed below, and few current books teach three acts to novelists instead of screenwriters. He lands about where this explanation lands. If you carry screenplay numbers into a hundred-thousand-word novel, remember that they were measured on something else.

Knowing the objections makes you better with the model. None of this settles whether you should use three acts. A writer who knows all five objections uses the model better than one who has only been told Aristotle made it.

Real-world examples

Screenplay (1979): the one-page diagram behind the method

Field's book is the origin document of the prescription, so judge it by what it says and not by its reputation. The book is listed below. The whole claim rests on a single-page diagram, the Paradigm, which you can still download from his site. The diagram labels the acts set-up, confrontation and resolution. It splits Act Two into two halves, each with its own dramatic context. For a 120-page script it puts Plot Point I roughly at pages 20 to 30, Plot Point II roughly at pages 80 to 90, and the midpoint near page 60.

The method started as a reader's problem. While he ran the story department at Cinemobile Systems in the early 1970s, Field read roughly two thousand scripts and recommended roughly forty. If you turn down that many, you need a theory of what you're looking for, and the paradigm is that theory written down. So the paradigm works better as a checking tool than as a creative one. Field built it to give a quick verdict on whether a script had structure. He didn't build it to tell a writer what to write.

The paradigm hands novelists two blind spots. Its unit, the screenplay page, has no equivalent in prose. And it measures position, which is the dullest thing about a turning point. Field himself was milder than his image. A ten-page window is no rule. The tightening came as the model spread, and it didn't come from him.

Poetics and Die Technik des Dramas (fourth century BC and 1863): what the sources say

Two primary sources sit under the history of three acts. Both are free online, and both are often misreported. In the Poetics, at 1450b, Aristotle says a whole has a beginning, a middle and an end, each defined by cause. He's giving you a test for whether a plot is complete. He isn't offering a method for splitting a play into performed sections, and the word "act" doesn't appear. At 1452b he gives his real division of tragedy: prologue, episode, exode and choral song, plus the two-part split into complication and unraveling. Nowhere does he lay out three sections with positions. The Climax page works from the same text. Aristotle has no word for the modern climax, and the diagram with its peak in the middle, so often credited to him, is Freytag's.

Freytag's book from 1863 has five parts, with the climax in the middle and a catastrophe at the close. In a tragedy, the catastrophe means most of the cast ends up dead. The Resolution page shows this using Elias J. MacEwan's English translation, which is free at the Internet Archive.

In practice, the two misreadings lead to two bad drafts. If you believe Aristotle prescribed three acts, you treat the model as beyond argument and obey the percentages instead of the story. If you absorbed the school pyramid, you put the big event in the middle of the book and then spend roughly forty percent of the novel on the aftermath. Freytag's shape does fit that plan, but he was describing five-act verse tragedy. Mix your sources, taking the shape from one and the proportions from the other, and you end up with a resolution that runs eighty pages.

Northern Lights (1995): three printed parts you can check

Philip Pullman's Northern Lights opens His Dark Materials. Scholastic published it in the UK in 1995, and in the US the book is called The Golden Compass. The novel is printed in three named parts: Oxford, Bolvangar and Svalbard. Because the divisions are printed, you can check the mapping for yourself.

Oxford is act one in the strict sense. The part sets up Lyra, her college, the rules of her world and the children who keep vanishing, and it ends with her leaving. To be exact, it ends once going back to Jordan College and her wild life on its roofs is closed to her. Here the act-break question earns its keep. The Gobblers' inciting business falls well inside the part. The break comes later, when her old life shuts.

Bolvangar is act two, and it matches the model. The part is a chain of attempts, and each one shifts the situation without settling it. Lyra goes through Trollesund and a witch-consul, an armored bear, a trek north, a lost boy, and finally the station. Her assets climb while her safety falls, and both of those directions are right for a middle act.

Svalbard is act three. No options are left, the showdown is forced, and the finish rests on the result of one decision. Even so, the book doesn't fully resolve. The ending shuts one question and deliberately leaves a larger one open, because the novel is the first of a trilogy.

The ending shows the series strain clearly. Three acts describes how each volume is built and says nothing about the whole set. Plan a series with three acts as your entire plan and your first book will either resolve too much or fail to stand on its own. The job that runs across volumes belongs to Series Arc.

Murder on the Orient Express (1934): three parts with a different engine

Agatha Christie's novel also comes in three parts. The split is real, but the labels don't match the model: "The Facts," "The Evidence," and "Hercule Poirot Sits Back and Thinks." The same book turns up on the Setting page, for the snowdrift that caps the number of suspects, and on Maps, for its printed plan of the sleeping car. Three pages lean on how neatly the book is built. Here only the split into parts matters.

Part one does act one's job. You get the train, the passengers, the snowdrift and the corpse. Part two is interviews, one passenger after another. There's no confrontation in it and no escalation of the act-two kind. Instead the facts pile up. Each interview supplies a fact and resolves nothing, and the pressure on you as a reader comes from the growing heap of facts. Nobody is being cornered. Part three is thinking. Poirot goes back over what he has, and then the solution follows.

A mystery's middle runs on a different motor. In a typical act two the reader asks "will she manage it." In a mystery's middle the reader asks "what is actually the case." A satisfying middle releases information in a controlled order. You can squeeze the book into three acts, and plenty of writing on structure does. But then you have to call the interviews confrontation, and that label hides how the section really works. For crime writers, the general model helps least here and Mystery helps most.

Storytelling in the New Hollywood (1999) and Into the Woods (2013): two repairs to act two

Kristin Thompson and John Yorke made the two most serious published repairs to act two. They came at it from opposite sides and ended up almost in the same place.

Thompson worked from a sample of films and divided them into four sections of about equal length: setup, complicating action, development, and climax plus epilogue. She says outright that she did so because the three-act model fails to describe how Hollywood films are built. Her second and third sections are Field's confrontation, split at the point most writers call the midpoint. Five of the films in her sample are Tootsie, Back to the Future, Groundhog Day, Amadeus and Alien, so you can check her claim against films you can watch tonight.

Yorke's Into the Woods was first published in 2013, with a US edition from Overlook Press in 2014. He wrote it out of a long career running television drama, and he argues for five acts drawn from Shakespeare's practice. The subtitle reads like an attack on three acts, but his position is milder. He calls five acts three acts refined, and he says three acts are the basic model under almost every successful story. The two extra breaks buy control. You get five divisions instead of three, with turning points spaced evenly enough that tension doesn't drain out of the middle. His own site has a free summary that works the idea through Macbeth and The Godfather, and his book is listed below.

Put together, the two teach one lesson. Keep three acts as the container, because it travels well and is hard to misuse. Then add at least one planned break inside act two, which is the Midpoint's job. The missing break explains why writers who hold only three acts write a middle that runs out of steam at around forty percent.

An everyday scenario: the draft whose act one ran two hundred pages

Here's a made-up case. A novel draft follows Win Caddick, who keeps bees. Her hives stand along a field boundary, and the farm on the other side, Lowdham's, sprays.

The author shows the draft to her writing group. The manuscript runs four hundred pages. Pages 1 to 200 cover the death of Win's mother, Win inheriting the land, two summers of beekeeping, a failed relationship, a new farm manager arriving at Lowdham's, and plenty of weather. On page 203, a hive dies. The group's verdict is that the middle sags. In fact, act one is two hundred pages long and the middle hasn't started yet.

Run the act-break question scene by scene. After her mother's death, is Win's old situation still open to her? Yes. After the inheritance? Yes. After the first dead hive? Still yes. Then she collects the dead bees in a bag, drives them to a lab and pays for an analysis she can't afford. A positive result will force her to accuse a neighbor she has to live beside for the rest of her life. Now her old situation is closed, and that scene is the act break.

Everything before it is setup. Act one takes half the book, where the usual share is a quarter, and the setup this story needs would fit in about a third of those pages. Two hundred pages is about three times too much.

The second finding is separate. Act two turns out to be one move tried four times. Win politely asks Lowdham's to stop spraying, and each time they politely refuse. Four refusals of the same kind are one event written four times. The repair here isn't structural. The fix comes from the Try-Fail Cycle and its rule that the outcomes have to differ. Win appeals to the farm, then to the regulator, then to the insurer, then to the local newspaper. Each one is a different closed door, and she holds less after every one. Once the only thing left is the thing she swore she would never do, act three starts on its own.

Three acts found the fault in about twenty minutes, working from a chapter list. On how to repair it, the model had nothing to say.

Try it yourself

You need your draft, or a detailed synopsis, and somewhere to write a chapter list.

  1. List what changes in each chapter. Write one line per chapter: the chapter number and what changes in it, not what happens. A line like "they travel to the city and talk about the plan" tells you nothing changed, and you already have a finding.
  2. Find the first act break with the act-break question. Go chapter by chapter and ask whether the protagonist's earlier situation is still open to them. The first chapter where the honest answer is no ends act one, wherever it falls.
  3. Measure once, then stop. Work out what share of the book act one takes up. Above roughly a third, you have a setup problem. Below roughly a tenth, readers may not yet care what is being lost. Treat the number as a warning light and never as a goal, and don't move scenes around to hit a figure.
  4. Find the second break the same way. Ask whether anything is left except the hardest option. The first chapter where every easier option is shut opens act three. Where the biggest explosion happens doesn't matter.
  5. Audit act two with one column. For each chapter in the middle, write down what the protagonist loses: options, allies, information, time, standing, money or self-respect. A run of chapters with nothing in the column means sag. The same loss over and over means one chapter written several times. Losses that shrink toward the end of act two mean your middle is running in reverse.
  6. Check that the win in act three was bought in act two. Name the asset that wins the climax. The asset might be a skill, some knowledge, an ally, or a choice the character couldn't have made in chapter three. Then find the chapter in act two where it was bought. If there isn't one, you have a rescue ending and an act two without a spine.
  7. Check that act three answers act one. Write the Dramatic Question in one sentence, in the form "will X manage to Y." Then confirm that the Climax answers it with yes, no, or yes at a price. Asking one thing and answering another is the top cause of readers saying an ending came "from nowhere."
  8. Test the list against another shape on purpose. Run the same chapter list against Kishōtenketsu and ask where the turn that changes everything lands, or against Seven-Point Structure and ask whether the pinch points land anywhere. You aren't hunting for a better template. You want to learn whether your three-act reading rested on evidence or on assumption. Give this step about twenty minutes.

Common mistakes and misunderstandings

Crediting Aristotle with three acts. You'll hear it all the time: "Aristotle invented the three-act structure." No error about structure gets repeated more often. The Poetics says a whole plot has a beginning, a middle and an end, defined by cause. The book has no acts, no positions and no three-way split for performance. The oldest surviving rule on the number of acts is Horace's, and he says five. The three-act method is Field's, from 1979.

Reading Freytag's pyramid as three acts with new names. His pyramid has five parts, a climax in the middle and a catastrophe at the close, and its subject is five-act verse tragedy. Mix its shape with modern proportions, or the other way around, and you get an eighty-page resolution. See Freytag's Pyramid and Resolution.

Treating percentages as targets. If a finished draft has a beat at 23 percent, you've made an observation. Moving a scene so it lands at 23 percent of an unfinished draft is superstition. Field's windows were generous, and prose has nothing to match the page unit his numbers rest on.

Assuming act two simply gets written. Act two has to be filled. Writers who hold only three acts get through act one and then improvise two hundred pages. Act two needs its own equipment: attempts that escalate and turn out differently, a midpoint that changes the terms, and subplots that cost something. Without that equipment, "act two" is only a name for the place where the trouble starts.

Calling the model neutral, or throwing it out once you learn it isn't. Salesses argues that a causal, desire-led, three-act shape is one tradition's expectation. His claim is serious and deserves an answer. The model still serves a writer aiming at Anglophone readers who expect it. Keep both truths in your head at once.

Giving the first book of a series a pass. An ending that settles nothing gets no exemption. The container governs each volume, and the larger question belongs to Series Arc. Northern Lights closes its act three and opens its series question in the same last chapter, and you should aim for the same. "It's book one" doesn't excuse an ending that settles nothing.

Expecting a verdict on quality. Three acts is a structural test and nothing more. A draft can pass every question and still fail, with dull prose, a passive protagonist, or nobody able to say what the book is about. Structure gets overrated because it's the easiest thing to inspect.

Related elements