Back to the table

61 Structure Ci Story Circle Story Structures

Story Circle

Dan Harmon's eight-step loop.

What it is

Draw a circle. Cut it in half top to bottom, then again side to side. You now have four points where the lines meet the rim and four arcs between them. Those are the eight positions, numbered clockwise from the top, with the odd numbers on the crossings and the even numbers in the quadrants. That's the whole apparatus. The shape itself makes the case, because a circle gives you four things a line doesn't.

The top half is the world your character already lives in, and the bottom half is somewhere else. Harmon shaded the lower half on his own diagram and labeled the two halves the ordinary world and the special world. Everything between the three o'clock crossing and the nine o'clock crossing happens away from home, under different rules.

The story goes down and then back up. In the first tutorial, Harmon calls it a pattern of descent and return, diving and emerging. The second tutorial makes the metaphor explicit and slightly mad. Your mind is a house, consciousness is the swept upstairs, and the unconscious is the basement with the fuse box in it, so a story is a trip down there and back for something you need. You can take or leave the psychology. What matters for structure is that the lowest point of the circle, position five, is a finding rather than a disaster: the thing the character came for, reached at the furthest distance from home.

Positions face each other, which a line can't give you, and it's the part most useful to steal. One sits opposite five, so the comfort your character starts in is answered by the thing they find at the bottom. Two sits opposite six, so what they wanted is answered by what it costs. Three sits opposite seven, so the way out is answered by the way back. Four sits opposite eight, so whatever they learned while adapting is what they use at the end. Harmon spells this out in the fourth tutorial, including the practical version: when the ending isn't working, look at the position directly across the circle from it.

Position one and position eight are the same place. Not a similar place. The same kitchen, the same reef, the same apartment, the same job, with a different person in it. The circle only closes if the reader can stand in the opening scene and the closing scene and see what moved.

The familiar beats fit onto the circle roughly like this. The Inciting Incident lives around position two, and the First Plot Point is the crossing at position three. The Midpoint and position five describe the same hinge from different angles. All Is Lost usually falls in the costly stretch at position six, and positions seven and eight contain the Climax and the Resolution. What this page covers is the loop they sit on. This table writes up the beats one at a time, so this page won't explain them again.

Dan Harmon is a television writer. He created Community, which began on NBC in 2009, and co-created Rick and Morty with Justin Roiland, which began on Adult Swim on December 2, 2013. Neither of those is a novel. The Circle is a television tool that novelists have adopted, not a novelist's tool that happens to work on television.

The circles began as private doodling. Harmon has said he started drawing them in the late 1990s while stuck on a screenplay. Brian Raftery's 2011 Wired profile of him describes circles covering the whiteboards and walls of his office. It also reports Harmon's own name for them, embryos, because each one contains everything a satisfying story needs. By then he was using them to lay out everything on Community, from a single joke to a whole season.

The written-out, numbered, teachable version came from somewhere more specific. In 2003 Harmon and Rob Schrab set up Channel 101 in Los Angeles, a monthly screening with two rules, more or less. A pilot or an episode runs five minutes or less, and the audience votes. The shows with the most votes are renewed for next month, and everything else is canceled. Harmon wrote the Story Structure tutorials for people submitting to that festival, and the format they were written for explains every design decision in the model.

Think about what Channel 101 demands. You have five minutes, and in them you must deliver a complete, satisfying story to a room, in front of the competition. If it lands, you have to do it again in four weeks with the same characters. If it doesn't, you're finished. A twelve-stage template with a mentor, a threshold guardian and a refusal of the call is useless at that scale. You need the smallest number of positions that still gives the feeling of a whole story, arranged so you can run them again next month. Eight positions on a loop is the answer to that brief.

You need to be careful with the sourcing, because it's messier than the craft blogs suggest. Harmon published the series as numbered tutorials on Channel 101's own site. They begin with Story Structure 101: Super Basic Shit and run through 102: Pure, Boring Theory, 103: Let's Simplify Before Moving On, 104: The Juicy Details, 105: How TV is Different and 106: Five Minute Pilots. At the time of writing, the Channel 101 wiki that reproduces them returns an HTTP error instead of a page. The texts described here come from archived copies of that wiki, which carry Harmon's byline, and the resources below link to the archive, not the live site. (An earlier page on this table had to drop a terminology point because of that dead wiki.)

On dates, the fair answer is a range rather than a year, and it isn't the year you'll usually see. The tutorials carry no publication date. Clues inside them put them in the mid-2000s. In the second installment, Harmon brags about having gotten ten episodes out of his own Channel 101 show Laser Fart, which dates from 2004. In the fifth, he names the NBC sitcom Happy Family as the only television he was watching, and it aired only between September 2003 and April 2004. The sixth installment sits in Channel 101's own reverse-chronological article index, between site news items from July and November 2005. So the Circle was drawn in the late 1990s, written up for Channel 101 around 2004 and 2005, and made famous later by Community and Rick and Morty. Anything that dates the Circle to 2009 is dating its audience, not its publication.

In Harmon's own words, the archived primary source does establish the geometry, the two halves and the opposing positions. It records his insistence that the structure itself is simply circular, and that any specific description of it belongs to your story rather than to him. It shows his flat refusal to hedge. He writes that he isn't going to keep saying there are exceptions, because exceptions are style, not structure. What it doesn't establish is a date, a copyright position, or any claim by Harmon to have invented the underlying pattern.

The Circle is a simplification of the monomyth, and Harmon says so himself. He isn't coy about it. The fourth tutorial carries a diagram of Campbell's stages laid over the circle. Harmon's caption notes that Joseph Campbell was a comparative mythologist rather than a screenwriting guru, and that this is where he, Harmon, thinks Campbell's chapters would fall if you forced them onto his circle. In the body of the same tutorial, he borrows Campbell's stage names out loud. The road of trials is the adapting stretch, and the meeting with the goddess is the bottom of the circle. The atonement with the father is the costly climb back, and the magic flight and the rescue from without are the return. He also names Christopher Vogler's and Syd Field's labels for the same places, then says which he prefers and why. Hero's Journey covers Campbell and Vogler, the record of what each of them actually wrote, and the long argument about whether the pattern is universal, so this page won't go over it again.

The question here is narrower. Why would anyone want a simplification, and what does it buy you? It buys you portability. You can hold eight positions in your head, draw them on a napkin, and apply them to a five-minute film, a chapter, a subplot or a trilogy, which you can't do with seventeen stages. It buys neutrality. The mythic trappings are gone, so there's no mentor you feel you must supply and no temple you feel you must build, and the model stops hinting that your story should be about a young man with a sword. And it buys symmetry: the opposing-positions idea above, which isn't in Campbell in that form and is the one truly new thing on the diagram.

What it loses is everything the extra stages were carrying. Harmon's single word for the whole adapting stretch covers what Vogler splits across several stages. On a five-minute film, that's an advantage. On a hundred thousand words it's a hole, and that hole is the subject of half the rest of this page.

Why it matters

It closes, which a list does not. This is the practical case, and the only one you need. A Beat Sheet is a list, and you can satisfy a list. You can write something at every line, tick each one, and still end up with a draft in which nothing has happened to anybody. A list is checked item by item, and no item on it has to refer to any other. Three-Act Structure has the same weakness in a milder form. A beginning, a middle and an end can all be there and still not be about the same person. You can't satisfy a circle item by item, because it has to meet itself. If position eight is the same place as position one, then to write position eight you have to know what position one was, in detail, as a state of affairs with a person standing in it. If position six is opposite position two, then the price has to be a price for the thing wanted, and a bad afternoon won't do. The shape forces the comparisons that a list leaves optional.

It works best as a diagnostic for the return, not as a plan. It asks three questions a linear template never makes you answer. What changed? Skip what happened and name what's different in the world at the end. In whom? Name them. It's usually one person. If the truthful answer is "the situation changed but nobody did", you have a plot and not yet a story (see Character Arc). At what cost? Position six exists because the model doesn't believe in free change. If nothing was paid, the reader will feel the ending as something you decided, not something that happened. That's a twenty-minute test on a finished draft, and it catches a failure that structural checklists are bad at: the book that hits every beat and still ends with a shrug.

The same-room test runs the closing question on your own manuscript. Start by finding the most concrete thing in your opening. Skip the theme and the want, and look for a thing: a room, a vehicle, a job, a weekly obligation, a person who always calls on Sundays. It should be something your character can physically walk back into. Then write the scene, or just the paragraph, in which your protagonist meets that same thing after the story is over. There are three possible outcomes, and whichever you get, you learn something. They behave differently, and you can say which position of the circle bought the difference. Then the circle is closed. Go check that the position you named is actually on the page and not just in your notes. Or they behave differently, and you can't say why. Then the change is only claimed. That's the most common result, and the repair is at position four or six. Either they never adapted to anything, or the adapting cost them nothing, so the ending has nothing to be made of. Or they behave the same. Either you've written an episode rather than a novel, which is a legitimate shape and is dealt with below, or the book has no eighth position and the ending is just where it stops. The test has an obvious limit. Heroine's Journey and the flat-arc tradition press on the same one. Some protagonists aren't supposed to change, and exist to change everyone around them. For those, run the test on a secondary character and on the room itself. If literally nobody in the closing scene is standing differently, the circle is open, whoever the protagonist is.

Its native habitat is the episode, and novelists skip past that fact more often than any other. The Circle was built to be run again next week with the same characters. Harmon is explicit about it. The fifth tutorial says television is no different in structure, but different in one practical sense. A film sends you out of the theater on a high. Television's job is to keep you watching for the rest of your life. Within an episode you still run the full loop. The emphasis of the eighth position changes: less stirring things up, he writes, and more getting things back to where they started. Then he gives a second, sourer version of the eight steps for television, in which the final position is the character admitting that change was futile. And he notes the obvious commercial reason. If the Death Star blows up, the series is over. John Yorke, who has run drama for British broadcasters, arrives at the same place from inside the industry in Into the Woods: A Five-Act Journey Into Story (2013; US edition Overlook Press, 2014, now published by Abrams Press). His own five parts are Home, Woodland, The Forest, The Road Back and Home Again, Changed, which is the circle written out at book length. His treatment of series drama argues flatly that the world of series is a world without change. In that world, each week the threat arrives and is expelled, and the family is restored. Two practitioners, working independently, describe the same reset.

For a novelist that leaves a pair of lessons. The first is that the Circle is excellent at the single unit and vague about the whole. You might write in installments: a chapter, a novella, an episode of a serial, a book in a long Series. If so, the loop is a really good shape for the installment, because it makes each one complete in itself, and that's what keeps a reader buying the next. What it won't do is manage the thing that must not reset across installments. That's the subject of Series Arc, and the Circle has nothing to say about it. The second is that a novel ends for good, like a canceled show rather than a renewed one. The whole money logic of the television version (restore the situation so we can sell it again) is the opposite of what a novel's ending has to do. Say you take the Circle from television and leave its eighth position shaped the way television shapes it. You'll write a novel whose last chapter puts everything back, and readers call this "nothing happened". Keep the loop. Refuse the reset.

The loop strains in five places.

Three charges have been made against it, and this page rules on none of them. Each is given at its strongest and credited to the person who made it.

The first charge is that the Circle is the hero's journey narrowed, and the narrowing throws away real options. The script editor and author Lucy V. Hay made this case directly in a post on her Bang2write site on January 6, 2020. She doesn't object that Harmon hid the debt, because he didn't. Her objection is that the Circle inherits one plot archetype, the voyage out and back, and treats it as the shape of story itself. There are other long-recognized patterns, rebirth among them, that it doesn't describe. She also presses the flat-arc point: not every character changes, since some are change agents, and the Circle's eighth position doesn't allow for that. Her conclusion is that it should be one tool for looking at structure, not the only one. If you want the longer version of the "there are several of these, not one" argument, see Seven Basic Plots.

The second charge is that the universality claim is the weak part, and that as stated it's unfalsifiable. Nothing could ever prove it wrong. Harmon's first tutorial tells you to look at your favorite films, your party anecdotes, your most vivid dreams and the shape of a pop song. It tells you to see the pattern everywhere, and to trust that the pattern is hardwired into your nervous system. Something that can be found everywhere can't be tested anywhere. The strongest published version of the objection to structures of this kind is Oren Ashkenazi's, in a Mythcreants piece of March 15, 2025. His charge is that models like this set arbitrary requirements, and he names the Circle's third position, the move to an unfamiliar situation, as his example. When a structure is sold as an all-in-one account of how stories work, he argues, writers following it have to throw out every idea that doesn't fit, which will be most of their ideas. The academic form of the same objection is a conference paper by Hartmut Koenitz, Andrea Di Pastena, Dennis Jansen, Brian de Lint and Amanda Moss, "The Myth of 'Universal' Narrative Models" (ICIDS 2018). It catalogs narrative structures from outside the Western print and cinema canon, to argue that the monomyth-and-dramatic-arc family isn't the universal model it's taken for. Harmon has pushed back on the reading rather than on the model. In a pandemic-era interview with Ari Melber on MSNBC, reported by Jason Hellerman at No Film School in June 2021, he says people mistake him for claiming that all stories are the same. The claim he's actually making, he says, is the subtler one that information has shapes. Kishōtenketsu is this table's page for the strongest counter-case, and Hero's Journey carries the full universality dispute with its named critics.

The third charge is that the Circle became visible, which is a problem for a method that works by feel. It's now among the most widely taught structures on the internet, and the two series it ran are among the most heavily analyzed on television. A large audience has learned to see the scaffolding. The sharpest documented case comes from inside Harmon's own show, and it's the third example below. At its strongest, the charge runs like this: a structural model whose power depends on an audience feeling the shape, without noticing it, is damaged by becoming famous.

Real-world examples

Matt Haig's The Midnight Library (2020): position one and position eight in the same flat

Haig's novel was published by Canongate in the UK in August 2020 and by Viking in the US. It's the cleanest closed circle in recent commercial fiction, and it makes a good demonstration because the circle is doing structural work, not decorating. What follows gives away the last movement of a very widely read book.

Position one is Nora Seed's life in Bedford, where almost nothing is going well. Position two is the want, and Haig is unusually blunt about it. She doesn't want a better life so much as she wants out of this one. Position three is the crossing, and it's a suicide attempt. Afterward she arrives in a library between living and dying, run by Mrs. Elm, who was her school librarian, where the volumes are the lives she might have had.

The adapting quadrant is where the novel earns its structure, as Nora learns the rules and then tries life after life. In one she's married to the man she left. In others she's a rock musician, an Olympic swimmer, a glaciologist, and a Cambridge philosopher with a husband and a child. In each one she picks up the skills on arrival and then finds the life wanting, and that's what position four is for.

Position five is the one she would keep, the life that looks like the answer from outside. Position six is the price, and it's the best-built thing in the book. She can't stay, because the self who would have lived that life isn't the self who arrived in it, and the library begins to come apart around her as she tries. Positions seven and eight are the return. They take place in the same flat, on the same night, in a life that hasn't changed in a single outward detail, and that's the reason to use this novel as the worked example. Nothing in the world is different. Everything in the person is. If you want proof that the eighth position is about who is standing in the room rather than what's in it, this is the book.

What a writer should take from it is that a comparison the reader can make is what closes the circle. You can measure the change because Haig puts you in that flat at the start and takes you back to it, deliberately, at the end.

Pixar's Finding Nemo (2003): a circle you can see from the air

Andrew Stanton's film, released on May 30, 2003, is useful here for the opposite reason. It's so easy to read that you can mark the positions with very little argument, which makes it a good place to watch the opposing positions at work.

Position one is Marlin's anemone on the edge of the reef. The opening sets up a comfort that is also a cage. The drop-off is right there, and the one thing he forbids is going near it. Position two is Nemo's capture. Position three is Marlin leaving the reef, which the first scene told us he couldn't do. The adapting quadrant is the journey out: Dory, the sharks, the trench, the jellyfish, the turtles, the whale. Each one is an adaptation, not a victory. Position five is Sydney Harbour and the fact of finding his son. Position six is what it costs. He believes he has lost Nemo, so the price is paid in the very thing he was trying to protect. Positions seven and eight bring them back to the reef, and the ending is Nemo going off to school and Marlin letting him go.

Now look across the circle. Position one was a father who wouldn't let his son near the drop-off, and position five is the father who crossed an ocean to reach him. Position two was the loss of the child. Position six is the loss apparently repeated, with the father now in a place to choose differently. Position four taught him to rely on someone unreliable, and position eight is him trusting a small fish with a weak fin. That reading doesn't ask you to squint, and it shows what the opposing positions are for. They're a way of making an ending refer to a beginning.

Notice the one thing the film doesn't do, which is reset. Marlin's anemone is the same anemone, and his relationship to it has changed. That's a novelist's eighth position rather than a television one, and it's useful to set it beside the series logic described above.

Rick and Morty, "Never Ricking Morty" (2020): the model drawn inside the story

When a method becomes part of popular culture, that's a real event in its history, and this episode is where it happened. Rick and Morty's season four episode "Never Ricking Morty" was written by Jeff Loveness and first broadcast on May 3, 2020. It's set on a train that generates stories, and the map of the train that the characters find is Harmon's circle. Tom Reimann reviewed it for Collider on May 4, 2020. He identified the map as the Story Circle, the template Harmon famously uses to plot his scripts. He described the episode as wholly meta and about the process of writing an episode of that very show. He judged that the quantity of scriptwriting inside baseball left it without a coherent storyline.

The lessons from it point in opposite directions. On one side, the Circle's own author thought it was funny and useful to show the audience the machine, which tells you something about how seriously he wants it taken as a law. On the other side is Reimann's criticism, which names the cost. Once the diagram is in the room, the episode is about the diagram. In prose, that's the novel whose structure you can feel being carried out. If a reader can see your positions, they're reading your plan.

An everyday scenario: the farrier who came home for nothing

Marsali Bowen is a farrier in Culverhay, in a draft of about ninety thousand words. Chapter one is her route: three barns, a riding school, and an old horse she'll have to tell someone about. In chapter six she takes a season's work at a big show barn two hundred miles away, because her own client list is shrinking, and she wants, just once, to shoe horses that matter. The middle of the novel is that barn. The last four chapters bring her back to Culverhay and the same three barns.

On paper the circle is closed. There's a comfort, a want, a crossing, two hundred pages of adapting, an achievement, a return and a homecoming. Yet the writing group said it was well observed, and that the ending was somehow flat.

The same-room test finds the hole in ten minutes. The concrete thing in chapter one is the riding school: a Tuesday morning, four ponies, and an owner who argues about the bill. Write the scene where she goes back. What does she do differently?

In the draft as it stands, nothing. She does the four ponies, and she takes the argument about the bill. When you ask which position of the circle would have bought a difference, you find the sixth position is missing entirely. At the show barn she was good, and she was noticed. She was offered a permanent job and turned it down to come home, which is a decision and not a price. Nothing was taken from her. The eighth position has nothing to be made of, and no amount of lyrical writing about the Tuesday morning light will supply it.

There are a couple of ways to fix it, and both of them repair position six, not the ending. The first is a cost inside her trade. At the big barn she signs off on a horse she had private doubts about, because the barn's vet and the barn's owner and the barn's schedule all say it's sound. Three weeks later it breaks down. Nothing is proved against her, and nobody blames her. Now the Tuesday morning at the riding school is a different scene. She picks up the fourth pony's foot and won't put it down, and the owner arguing about the bill is arguing with somebody who can no longer be hurried.

The second is a cost in Culverhay, which is better still, because it uses the two-and-six opposition. What she wanted in chapter six was to shoe horses that matter. Make the price this: while she was away, one of her three barns found another farrier and kept him. She comes back to a route with a hole in it that she dug herself. Position eight is then her standing in the riding school yard, knowing just what the season cost. The question the book has been asking (was it worth it?) gets an answer the reader can weigh, instead of an assurance.

Notice what the Circle did and didn't do here. It found the fault in about ten minutes from a chapter list, and it found it in the right place, which wasn't the ending. It had no opinion at all about which repair to choose.

Try it yourself

  1. Draw it, by hand, small. Draw a circle and two lines, and number the positions one to eight clockwise from the top, with the odd numbers on the crossings. Don't open a template. The point is that it fits on the back of an envelope and takes fifteen seconds, which is why you'll do it again.
  2. Fill position one with a place and a routine, not a mood. "Unhappy in her marriage" is not a position one. "Sunday lunch at her mother-in-law's, every week, for nine years" is. The whole model depends on your character being able to walk back into position one later.
  3. Write position two as one sentence with a verb they could act on. If you can't, you have a situation and not yet a story. Want vs Need has the fix. The difference matters here, because position five tends to deliver the want, while position six tends to expose the need.
  4. Name the crossing at position three and check that there's no going back. Ask the question Three-Act Structure uses for the same joint: after this scene, is the old life still available to them? If it is, they haven't crossed yet, and your bottom half hasn't started.
  5. Write position six before you write positions seven and eight. Of everything on this page, this instruction is the one that will change your draft. Put down, in one line, what the thing at the bottom of the circle costs. It might cost standing, money, a relationship, a belief, an ability or a person. Writers who plan the return first almost always write a return that was free.
  6. Run the same-room test. Take the concrete thing from your answer to step two, and write the scene in which your protagonist meets it after everything. Then say out loud which numbered position paid for the difference. If you can't name a position, the ending is only claimed.
  7. Check the length of the adapting quadrant. Count the words between position three and position five. In most novels it's thirty to forty percent of the book, and the Circle has just one word of advice about all of it. Go to Try-Fail Cycle and the Midpoint, and put real machinery in there. Treating position four as a single beat is how a middle sags.
  8. Run a circle for each viewpoint character, and one for each book if you're writing a series. An ensemble member whose circle has no sixth position is a passenger. A series where every volume resets to position one has brought in television's eighth position. See Series Arc for the thread that must not close.
  9. Run the loop on one chapter. The Circle can do this and longer templates can't, so it would be a waste not to use it. Take a chapter that readers skim, or that you dread rewriting, and mark its own eight positions. Where does the person start, and what do they want in these pages? What do they step into, and what do they have to work out? What do they get, and what does it cost them? Where do they end up, and what's different by the last line? A chapter with no sixth position is a chapter in which nothing was spent, and that's the most common reason a scene reads as filler. The diagram is small enough that you can do this to twenty chapters in an afternoon.

Common mistakes and misunderstandings

"It is just the hero's journey, so I can skip it." Harmon agrees about the ancestry and says so on his own diagram, so there's no concealment to expose. The simplification isn't a lossy copy of Campbell. It adds the two things Campbell's list doesn't have, a closed loop and a set of opposing positions, and those are the only reasons to use it. Dismiss it as a copy and you throw away the new parts.

Eight positions are not eight chapters, or eight scenes, or eight of anything. Four of the eight are arcs rather than points, and they aren't the same length as each other. In a novel, the adapting quadrant will be several times the size of the crossing. Anyone who hands you a page count or a percentage for each position has made it up. Harmon gives no numbers at all.

Don't put the price in the epilogue. The sixth position is a position, with scenes in it, roughly two thirds of the way around. If the cost of your story's central achievement turns up in the last chapter as a rueful paragraph, the reader has already finished weighing the book.

"My protagonist does not change, so the circle doesn't apply." This is the model's weak point, not yours, and Lucy V. Hay's 2020 objection is the one to read. The workaround is to run the eighth position on somebody else, or on the room. A flat-arc novel still has to show somebody standing differently at the end, or it hasn't closed. Heroine's Journey and Character Arc both bear on this.

The special world doesn't have to be a place. The lower half of the circle is a change of rules, and you don't have to change zip codes to get there. A marriage, a diagnosis, a secret, a new job in the same building, a correspondence: any of them can be the bottom of the circle. Writers who read the third position too literally end up inventing a journey their book didn't need.

Watch for the quadrant that's secretly a list. Your protagonist might do six interesting things in the middle. That's only an adapting quadrant if the six things are turning her into whoever has to exist at the bottom of the circle. The test is whether you could shuffle the order. If you could, you have episodes. On television that's a living. In a novel it's the sag.

A model that fits everything has told you nothing. The Circle will map onto anything you point it at, including a shopping trip, and Harmon's first tutorial actively encourages you to see it everywhere. That makes it a fine lens for describing a story and a poor proof of anything. When you find that your draft fits, you've learned that the draft isn't broken in that particular way. You haven't learned that it's good.

Reading it off a craft blog is reading a paraphrase of a paraphrase. The tutorials are short, free and archived, and they're written in Harmon's own very unbuttoned voice. The whole of the first one is about four hundred words. Secondary accounts routinely date the Circle to 2009, turn its positions into percentages, and credit Harmon with rules he never wrote. Go to the archive links below and read the man.

Related elements