Screenwriting

A sentence that cannot be photographed

Here is a sentence from a novel that does not exist: Marguerite had never forgiven her brother for the summer she turned nineteen, though she had long ago stopped saying so out loud. It is a decent sentence. It carries a grievance, a date, a silence, and two people. Now try to film it. Point a camera at a woman and record the fact that she has not forgiven someone for twenty years. You cannot. There is no lens for it. There is no lens for "had never forgiven," none for "long ago," and none for a thing somebody has stopped saying.

This is the whole problem of screenwriting compressed into one line, and everything else in this course grows out of it. A novelist can write interiority directly. A screenwriter can only write what a camera can record and a microphone can hear, and must arrange those recordable things so that an audience assembles the interiority for itself. Here is one solution, in format:

INT. FUNERAL HOME, VIEWING ROOM - DAY
A receiving line. MARGUERITE, 40s, coat still buttoned, works down it, shaking hands, saying the right things.
She reaches DANIEL, 40s. Same jaw, same eyes, twenty years of nothing between them.
She shakes his hand the way she shook the last three, and moves on.

Nothing there states the grievance. A viewer gets it in about two seconds, from one handshake given at the wrong temperature, and gets it more strongly than if a character had announced it. That transfer, from stated interiority to observable behaviour, is the craft. The format you are about to learn exists to hold the result.

Key idea: You may write only what a camera can record and a microphone can hear. Everything else must be given a physical form, or it does not exist in your film.

What the format is actually measuring

A screenplay is typed in Courier at twelve point, a monospaced typeface built for typewriters, at ten characters to the horizontal inch and six lines to the vertical inch, giving about fifty-five lines on a US letter page. Nobody chose Courier for beauty. It survives because monospacing makes a page hold a predictable quantity of speech and action, which makes page count a usable unit of time. One correctly formatted page runs roughly one minute of screen time.

Roughly is doing real work in that sentence, and it is worth knowing where the estimate is strong and where it is weak. A page of dialogue is close to reliable, because people speak at a fairly steady rate and the dialogue column has a fixed narrow width. A page of action is much less reliable: the single line "She opens the box" can occupy two seconds or ninety, depending on what the director decides is happening. So a hundred-and-ten page script is a hundred-and-ten minutes give or take about ten, and a script that is heavy on wordless action can run far longer than its page count suggests.

Why does anyone care? Because page count is the number every other department multiplies by. A producer prices a film partly from length. A first assistant director schedules by eighths of a page, dividing each page into eight parts and asking how many eighths a crew can shoot in a day, with two to five pages a day being ordinary on a modest feature. A financier reads length as running time and running time as how many screenings a cinema can fit into an evening. If you set your script in a proportional font at eleven point to squeeze it under a hundred and twenty pages, you have not shortened your film. You have broken the instrument everyone downstream is reading.

The margins follow from the same logic. Action runs the full width of the text block. Dialogue sits in a narrow column indented from the left, character cues sit further in and in capitals, parentheticals sit between the two. The precise figures vary by a tenth of an inch between style guides and software defaults, and no reader has ever rejected a script over a tenth of an inch. What matters is that the shape is instantly recognisable: a page of your script, seen from across a room, should be identifiable as a screenplay and nothing else.

The six elements

Almost everything on a screenplay page is one of six things. Learn them as jobs, not as indentations, because the indentation is what software does for you and the job is what you have to decide.

Element	Example	Its job
Scene heading (slugline)	INT. FUNERAL HOME, VIEWING ROOM - DAY	Interior or exterior, where, when. Drives the shooting schedule.
Action	She shakes his hand the way she shook the last three.	Present tense description of the recordable world.
Character cue	MARGUERITE	Names who speaks. Capitals, centred column.
Dialogue	You look well.	What is said.
Parenthetical	(not looking at him)	A brief delivery or action note. Use rarely.
Transition	CUT TO:	Editing instruction. Largely vestigial in modern scripts.

Three details are worth pinning down. First, a character's name is capitalised in the action line the first time that character appears, and normally thereafter written in ordinary case; major sound effects are capitalised too, which is a courtesy to whoever builds the sound. Second, cue extensions carry information: V.O. means voice over, the speaker is not physically present in the scene at all, while O.S. or O.C. means off screen, the speaker is present but out of frame. Confusing those two puts a sound recordist in the wrong room. Third, the word CONT'D after a cue means the same character is still speaking after an interrupting line of action.

The point: The elements are not decoration. Each one is a message to a specific department, and using the wrong one sends the wrong department the wrong information.

White space, and the twenty-second read

Sit with a professional reader for an afternoon and you will see something that no formatting guide mentions: they turn pages very fast, and they decide how they feel about your script long before they have finished it. A page of dense unbroken action, six or eight lines deep, reads as slow before a word of it has been processed. A page with air in it reads as quick. This is not fair and it is entirely real.

So treat white space as a craft variable you control. The working discipline is to keep action paragraphs to about four lines, and to break longer descriptions into separate paragraphs at the points where the eye would move anyway, which is usually where the camera would move or a new element enters. Compare:

Dense: Marguerite pushes through the double doors into the car park, where the rain has started, and she stands under the awning looking for her car among the rows of black vehicles, and behind her the doors swing open again and Daniel comes out holding two paper cups of coffee, one of which he offers her without saying anything.

Broken: Marguerite pushes out into the car park. Rain, just started.
She scans the rows of black cars. All of them look the same.
Behind her, the doors swing. DANIEL, two paper cups. He holds one out. Says nothing.

The second version is the same information. It also contains three implicit shots, a beat of comedy in the identical cars, and a silence you can feel. Line breaks in a screenplay are closer to line breaks in poetry than to paragraph breaks in prose: they control pace, and a reader hears them.

Master scene format and the shooting script

The format you have been learning is called master scene format. It describes what happens in a scene without specifying how it will be photographed. That is the format for a spec script, meaning a script you write speculatively, on your own, hoping to sell it or to get it made.

A shooting script is a different document, produced later, once a film is actually being made. Its scenes are numbered so the crew can refer to them, and it may carry camera and technical direction agreed between the director and the heads of department. Published screenplays are often shooting scripts, which is why beginners who learn format from published editions frequently arrive with scene numbers and camera angles in their spec drafts.

The practical rule for you: no scene numbers, and no camera direction, in a spec. Not because a rule forbids it, but because naming shots in a script nobody has agreed to make reads as a writer performing a job that has not been offered. Imply the shot instead. If you write that a hand closes on a key, you have written an insert without saying INSERT. Choosing what to describe is choosing the frame, and a good reader sees the film without your help.

Worth holding on to: Master scene format for a spec, shooting script for a production. Imply the shot by choosing what you describe, and let the director think it was their idea.

Unfilmables, and the honest exception

The rule that action lines carry only what a camera can record is the one you will break most, so it is worth stating the exception precisely rather than pretending there is none.

Pure interiority is out. You may not write that Marguerite remembers the summer of her nineteenth year, or that Daniel has always been the favourite. There is no image there, and an actor cannot play a fact the audience has not been given.

What you may do, in moderation, is write a line that is technically unfilmable but that tells a reader something a camera would communicate anyway. "DANIEL, 40s, the kind of man who irons his socks" cannot be photographed as written, but it tells the reader, the casting director and the costume designer exactly what to build, and the finished frame will carry it. That is a legitimate use. The test is whether the unfilmable line is doing the work of an image or standing in for one that you have not thought of yet. Writing "she is devastated" is the second kind, and it is a promissory note the film cannot cash.

The related argument concerns the phrase "we see." Some readers hate it as a needless intrusion of the writer. Others use it constantly. The defensible position is that "we see" is worth its space only when it makes a revelation land, as in a line where the point is the delay before the audience understands what they are looking at. Sprinkled through every third paragraph, it is filler.

Writing it, and protecting it

You do not need to buy software to write correctly formatted pages. Fountain is an open plain-text markup: you type in any text editor using simple conventions, a scene heading in capitals, a character name in capitals above the dialogue, and a converter renders standard pages. Trelby is a free dedicated program. Several browser-based tools have free tiers that will carry a short film comfortably. Final Draft is the industry default and costs money, and you will not need it until somebody else's workflow requires it.

On protection, keep two things separate in your head. Copyright in most countries arises automatically when the work is fixed in a tangible form, so your draft is protected the moment you save it, and in the United States formal registration with the Copyright Office is a separate step that strengthens what you can recover in a lawsuit. The Writers Guild of America registry is a different service again: it creates dated evidence that a specific document existed on a specific day, which is useful in a credit dispute, and it is not a substitute for copyright registration. Beginners routinely believe the guild registry is copyright. It is not.

Common misconceptions

• Format is gatekeeping. It is measurement. Page count feeds scheduling, budgeting and running time, so breaking the format breaks other people's arithmetic.
• One page always equals one minute. It is close for dialogue and unreliable for action. An action-heavy script routinely runs longer than its page count.
• Camera directions show the reader your vision. In a spec they show inexperience. Describe the thing you want framed and the framing follows.
• Published screenplays model what to submit. Many published editions are shooting scripts, with numbered scenes and technical direction that do not belong in a spec.
• Registering with the guild copyrights your script. It creates dated evidence of existence. Copyright is separate and, in most jurisdictions, automatic on fixation.
• Long descriptive paragraphs show craft. They read as slow. Four lines is a working ceiling, and line breaks control pace the way they do in verse.

Where this leaves us

• Only what a camera records and a microphone hears goes on the page; interiority must be converted into behaviour, object or line.
• Courier twelve point at ten characters to the inch makes page count a shared unit of time and money, reliable for dialogue and loose for action.
• Six elements do the work: scene heading, action, character cue, dialogue, parenthetical, transition, each addressed to a different department.
• V.O. means the speaker is not in the scene; O.S. means present but out of frame.
• White space is a pacing tool. Keep action paragraphs near four lines and break where the eye would move.
• Write specs in master scene format with no scene numbers and no camera direction; shooting scripts are a later document made by other people.
• Unfilmable lines earn their place only when they do the work of an image; "she is devastated" does not.
• Fountain and Trelby are free. Guild registration is dated evidence, not copyright.

Sources

1. Fountain. (n.d.). Fountain: A markup language for screenwriting. fountain.io
2. Writers Guild of America West. (n.d.). Writers Guild of America West. WGAW. wga.org
3. Wikipedia contributors. (2024). Screenplay. Wikipedia. en.wikipedia.org
4. Wikipedia contributors. (2024). Courier (typeface). Wikipedia. en.wikipedia.org
5. Cole, Hillis R., and Judith H. Haag. The Complete Guide to Standard Script Formats, Part 1: The Screenplay. Rev. ed., CMC Publishing, 1996.

Two ideas, only one of which is a film

In December 1982 a comedy opened in which an unemployable New York actor, so difficult that no director will hire him, puts on a dress, auditions as a woman, and gets cast in a daytime soap opera. Tootsie, directed by Sydney Pollack, with Dustin Hoffman as Michael Dorsey and Dorothy Michaels, and Jessica Lange in the role that won her an Academy Award. Hold that beside a second idea: a film about loneliness in modern cities.

Both are ideas. Only one of them is a premise. The difference is not quality, and it is certainly not that the second is a worse subject. It is that the first contains a person who wants a specific thing, an obstacle built into the want, a lie that has to be maintained, and a clock, because a soap opera shoots every week and the deception cannot last. The second contains a topic. You can film Tootsie. You cannot film loneliness, though you can absolutely make a film that is about loneliness, by first finding somebody who wants something.

This lesson is a filter. Most unfinished scripts are unfinished because the writer started from an idea that had no engine, discovered the problem around page forty, and blamed their discipline. The work below takes fifteen minutes per idea and saves you months.

The point: A subject is not a premise. A premise contains a person, a want, an obstacle and a reason it is happening now.

Idea, premise, logline: three different objects

These three words get used interchangeably by people who should know better, and separating them will make the rest of the course easier.

An idea is raw material: an image, a fact, a character, a question, a feeling, a news story. "Bees are dying." "My aunt sold her land." "What if you could not lie for a week?" An idea has no obligation to be dramatic. Collect them constantly and judge them never.

A premise is the idea plus the machinery that makes it move: who, what they want, what is in the way, and what it costs to fail. It is a working document for you, not a sales tool. It can be two or three sentences.

A logline is a premise compressed for someone else, usually one sentence, written to make a reader want to read the script. It is a marketing artefact and it comes last, though writing one early is a useful diagnostic because a premise that will not compress is often a premise that has not been built.

Take one idea through all three, so the difference is concrete.

Stage	The thing itself
Idea	My aunt kept bees and the hives died the year she did.
Premise	A woman who has not spoken to her father in nine years inherits his failing apiary and the debt attached to it. She has three weeks before the bank takes the land, and the only person who knows how to save the hives is the neighbour her father cheated. She wants the money from the sale. She needs to find out what actually happened between the two men.
Logline	Estranged from her beekeeper father until his death, a woman has three weeks to save his collapsing hives, with help from the neighbour he ruined.

Notice what happened between rows. The idea named a memory. The premise added a want (the money), an obstacle (the hives, the debt, the neighbour), a clock (three weeks) and a need distinct from the want (the truth about her father). The logline threw away the need, because a need is what the film discovers and a logline should not spoil it.

The dramatisability test

Five questions. Ask them out loud. If you cannot answer one in a single clear sentence, that is where the premise is broken.

1. Who wants what? A specific person, and a want you can state in about five words. Not "to be happy." Something you could photograph someone achieving.
2. What stops them? Opposition with its own reasons. If the only obstacle is that the protagonist is sad, you have a mood, not a story.
3. What happens if they fail? Consequences that are concrete and that the audience will believe. Losing the farm is a stake. Personal growth is not.
4. Why now? Something has changed to start the clock. Films almost never begin on an ordinary Tuesday; they begin on the Tuesday something arrives.
5. Can I see it? Does pursuing the want generate physical action, in places, with other people, that a camera could record?

Apply it to four candidates and watch three of them fail in different ways.

Candidate	Verdict	Where it breaks or holds
A film about loneliness in cities	Not yet a premise	No person, no want, no clock. It is a theme, and a theme belongs underneath a premise, not instead of one.
A biopic of a jazz pianist	Not yet a premise	A life is not a story. Ask which eighteen months, what he wanted in them, and who refused it to him. Then you have something.
A woman has three weeks to save her dead father's hives or lose the land	Holds	Person, want, obstacle, stake, clock, and the work is physical: smoke, frames, mites, a bank, a neighbour.
A night nurse finds a colleague stealing morphine	Holds, with a fix	Person and obstacle are there. The want needs stating: she wants to keep her job and her shift cover, which is exactly what reporting him destroys. Now the obstacle is inside the want.

Why this matters: The strongest premises put the obstacle inside the want, so that getting the thing costs the thing. That is not a rule of taste. It is what makes a second act possible.

Where premises actually come from

Nobody sits down and thinks of a premise. They arrive by a handful of routes, and knowing the routes lets you go looking.

The obsession. Something you cannot stop reading about. This is the most reliable source, because you will still be interested on draft six, which is the only version of the script that matters.

The what-if. Take a normal situation and change one rule. What if the person who has to identify the body is the person who reported them missing, and they are lying about both? What-ifs are cheap to generate and most are worthless, so generate twenty.

The collision. Put two things together that do not usually meet. A hospice and a heist. A wedding planner and a war crimes tribunal. Collisions produce tone as well as plot, which is why so many good comedies come from them.

The character in a box. Start with a person you can hear, then build the worst room to put them in. If your character is a man who cannot bear to be watched, the box is a job in which he is filmed all day.

The news item. Useful and legally hazardous. Facts are not copyrightable, but a specific published article's expression is, and identifiable living people have rights of privacy and, in some jurisdictions, publicity, plus the ever-present risk of defamation. If you want a particular person's story rather than the underlying events, that normally means acquiring life rights, which is a contract, not a courtesy. Public domain material and long-dead figures are the free end of this spectrum.

The engine test: does it generate scenes?

Here is the practical check that beats every abstract discussion. Take your premise and, in fifteen minutes, list twelve scenes it produces. Not a plot, not an order, just twelve situations in which somebody wants something from somebody else.

For the apiary premise: she signs for the coffin and the debt in the same hour. She opens a hive she does not know how to open and gets stung badly. She tries to sell to a developer who has seen the survey and knows the land is worthless. She goes to the neighbour and is refused. She reads her father's hive records and finds the neighbour's handwriting in them. She sells her car. The neighbour turns up at four in the morning with smoke and a spare veil and does not explain why. The bank sets a date. She finds the letter her father never sent. She learns the thing that makes the neighbour's version true. She has to choose between the money and the hives. She burns something.

Twelve came easily, and they are various: some are about money, some about knowledge, some about labour, some about the past. That variety is the engine working. Now try the same exercise with "a film about loneliness in cities" and you will produce four scenes, all of them a person alone in a room, and you will feel the machine seize.

A premise that generates twelve varied scenes in fifteen minutes will probably support a feature. One that generates five is a short film, which is not a failure, it is a form, and it gets its own lesson later. One that generates two is an image, and images belong in other people's films.

Two ways premises fail, and how each looks from the inside

Situation without story. The setup is vivid and nothing pushes. A family gathers for a funeral; secrets emerge. This can be a great film, but only once somebody in it wants something that another one of them will not give. The tell is that you can rearrange your scenes in any order without damage. If scene nine can trade places with scene four, no scene is causing the next one.

Story without pressure. The protagonist wants something and gets it too easily, or could simply walk away at any point. The tell is that you find yourself inventing obstacles that have nothing to do with the want: a car that will not start, a phone with no signal. Real obstacles come out of the premise. If your protagonist could solve the problem by leaving town, either take away the option or admit that the want is not strong enough.

There is a third failure that is harder to see, which is a premise that is really about the writer's opinion. If the answer to the question "what happens if the protagonist fails" is "then the audience will have learned that greed is bad," the film is an argument wearing a story. Arguments are allowed. They just have to lose sometimes, inside the film, to somebody the audience takes seriously.

Common misconceptions

• A strong theme is a strong premise. Themes have no clock. Put a person who wants something on top of the theme and the theme will take care of itself.
• Original means nobody has done it. Almost nothing is unprecedented. Originality lives in the specific handling: this character, this world, this pressure. Tootsie is a very old disguise plot handled precisely.
• You should not decide the ending before you write. You do not have to, but knowing what the want costs is not the same as spoiling your discovery, and it is what stops act two from wandering.
• Real events are free to use because they are facts. The facts are, but a living person's story usually is not, and life rights, privacy and defamation are all live risks.
• If I cannot write the logline, I need a better sentence. Nearly always you need a better premise. The sentence is a symptom.

Putting it together

• An idea is raw material, a premise is the machinery that makes it move, and a logline is the premise compressed for a stranger.
• Five questions decide dramatisability: who wants what, what stops them, what happens if they fail, why now, and can I see it.
• The strongest premises put the obstacle inside the want, so achieving the goal costs the thing the protagonist most needs.
• Want and need are different. The logline carries the want; the film discovers the need.
• Premises arrive through obsession, what-ifs, collisions, a character in a box, or the news, and the last of those carries rights risk.
• The engine test is twelve varied scenes in fifteen minutes. Five scenes is a short. Two is an image.
• Situation without story shows up as scenes that can be reordered; story without pressure shows up as invented obstacles unrelated to the want.

Sources

1. Wikipedia contributors. (2024). Tootsie. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Logline. Wikipedia. en.wikipedia.org
3. British Film Institute. (n.d.). Features and interviews. BFI. bfi.org.uk
4. Egri, Lajos. The Art of Dramatic Writing: Its Basis in the Creative Interpretation of Human Motives. Simon and Schuster, 1946.
5. McKee, Robert. Story: Substance, Structure, Style and the Principles of Screenwriting. ReganBooks, 1997.

A man stops his car for three horses

About ten minutes into Michael Clayton, written and directed by Tony Gilroy and released in 2007, the title character drives out of the city before dawn after a bad night. George Clooney's Clayton is a fixer at a large law firm: not a trial lawyer, a man who is sent when a client has done something that needs containing. He is in debt from a restaurant that failed, his brother is the reason it failed, and he has spent the night cleaning up after a client who ran somebody over.

On a hill, in the half light, he sees three horses standing in a field. He stops the car. He gets out. He walks up the hill toward them and stands there.

Then his car explodes behind him.

The plot function of the scene is obvious: an assassination fails because the target got out. But the character function is the thing to study. Clayton has, at that moment, no reason to stop. Nothing in his want, which is money and an exit, is served by walking up a hill to look at horses. He does it anyway, and he cannot explain it, and the film never explains it either. That unexplained act is the first hard evidence in the film that there is a person inside the fixer, which is exactly what the story is going to test. The horses are his need showing through his want.

Key idea: A character's want is what they pursue. Their need is what the story argues they actually lack. Keep them apart and the film has somewhere to go.

Want and need, and why they must not match

The want is external, concrete and pursued on purpose. You can photograph someone getting it. Clayton wants seventy-five thousand dollars and a way out from under his brother's debt. A want is the spine of the plot, because plot is what happens when somebody tries.

The need is internal, usually unadmitted, and is the film's actual argument. Clayton needs to stop being a man who is paid not to look at what he is holding. He does not say that, at any point. He would deny it if asked.

Now the crucial part. If want and need point the same way, there is no story: the character does the thing and is improved by doing it, and no choice is ever forced. The engine comes from the gap. The classic shape is that the protagonist can get the want or serve the need but not both, and somewhere in the last third the story presents the bill. This is why the ending of a well-built film feels inevitable and surprising at once: the choice was set up in the first ten minutes, in a scene that looked like something else.

Character	Want (external, stated)	Need (internal, argued)	Where they collide
The fixer	Money and an exit from the debt	To stop being paid not to look	The document that would clear the debt is the document that proves the client guilty
The beekeeper's daughter	To sell the land and clear the debt	To find out what her father actually did	The buyer needs the survey she has not read
The hospice volunteer	To get through his shifts unchanged	To stop treating other people as weather	The one patient who settles for him is the man who supplied him

Build the third column last. Writers usually know the want on day one and discover the need somewhere in draft two, and that is a normal order of operations, not a failure of planning.

Characterisation is not character

Characterisation is everything observable: age, job, accent, the car, the flat, the way she holds a phone, whether he salts his food before tasting it. It is the surface, and a surface is worth building because it is what a reader sees first and what an actor can use.

Character, in the older dramatic sense, is what is revealed when a person chooses under pressure. Give someone an easy choice and you learn nothing. Give them a choice where both options cost, and what they pick is who they are. This is the reason your protagonist must be put in situations where the honourable move is expensive, and it is the reason that a scene in which a character simply states their values is worth almost nothing.

A working consequence: whenever you catch yourself writing a scene that tells the audience what someone is like, convert it into a scene that makes them choose. A man who says he would do anything for his sister is a claim. A man who lies to a police officer for his sister and then cannot look at his own son is a character.

In short: Characterisation is what a person shows. Character is what a choice under pressure exposes. Write the choice.

Obstacles come in three ranges

Opposition is not one thing, and knowing which range you are working in tells you what kind of film you are writing.

Range	What opposes	What it produces	Risk
Inner	The character's own fear, addiction, self-image, grief	Interiority, intimacy, subtle films that depend on performance	Nothing visible happens; the film becomes a description of a mood
Personal	Other people with their own wants: family, lovers, colleagues, rivals	Scenes with two active parties, which is where dialogue lives	The world outside the room stops mattering
Social or physical	Institutions, law, money, weather, geography, war, machines	Scale, stakes, plot momentum	The protagonist becomes a tour guide through events

Most strong films operate at all three ranges at once, and align them. In the apiary premise, the inner obstacle is that she cannot forgive; the personal obstacle is the neighbour who will not help; the social obstacle is a bank with a date. When the three ranges pull in the same direction, a single scene can advance all of them, and that is what makes a script feel dense rather than long.

A diagnostic worth running on your own draft: take any ten pages and mark which range each obstacle belongs to. If every mark is inner, the pages will play as slow. If every mark is social, the pages will play as busy and unmoving. Both problems are usually fixed by adding the missing range rather than by rewriting the lines.

Building a character from behaviour

The standard advice is to write a biography. This mostly produces documents nobody reads, because a childhood is not playable. Here is a method that produces pages instead.

One. Give them a job with a verb. Not "she is a teacher" but "she stands in front of thirty people who do not want to be there and holds their attention for fifty minutes, five times a day." A job described as an activity generates behaviour immediately.

Two. Find the contradiction. Real people are inconsistent in patterned ways, and the pattern is the character. A paramedic who is calm at a road accident and cannot cope with a dinner party. A man who is generous with money and vicious about time. Write the contradiction down as a sentence with "but" in the middle, and you will find you can hear the person.

Three. Ask what they do when nobody is watching. This is where you find the small physical business that makes them specific: the way she reads the last page of a book first, the way he cleans a kitchen that is already clean at two in the morning.

Four. Ask what they would never do. Then find the pressure that would make them do it, because that is your third act.

Five. Write half a page of them talking, and throw it away. The point is not the page. It is to find out whether you can hear a voice yet. If every character sounds like you, you do not have characters, you have positions.

Worked example. Take the hospice volunteer. Job with a verb: he sits with people in the last four days of their lives and does not fill the silence. Contradiction: he is completely steady with the dying but cannot be in a room with his own mother for an hour. Alone: he counts things, doorframes, ceiling tiles, and does it without noticing. Never: he would never take a drink in front of another person, which is not the same as never drinking. Now you have a scene: his mother offers him a glass of wine at her table, and the whole of him is in what he does with his hands.

Arc, and the honest case against it

The received wisdom is that a protagonist must change. It is a good default and it is not a law, and the exceptions are common enough that you should know the shape of them.

A change arc takes a character who is wrong about something at the start and, through cost, moves them. Most mainstream drama runs on this.

A steadfast protagonist does not change; the world around them does, and the drama is in whether they hold. Long-running detectives, many westerns, and a great deal of comedy work this way. The audience is not asking whether she will become a different person but whether she will be broken. The pressure must be real or the film is a lecture.

A fall takes a character who is intact and shows the ways they consent to their own ruin, one plausible step at a time. The craft problem is making each step small enough to accept.

The genuine mistake is not choosing the wrong shape. It is announcing the shape: a character who says at the end that they have learned to trust people. If the events have not made the change legible, no line will fix it, and if the events have, the line is redundant.

Everybody else in the film

The single most useful test for a supporting character: what do they want from the protagonist, and what do they refuse them? A supporting character who exists to deliver information or to be encouraging is furniture. Give each of them a want that crosses the protagonist's want at least once and your middle section stops sagging, because the pressure now comes from several directions.

Second test: is there someone in this film who can say the thing nobody else can? Every strong ensemble has a character licensed by their position to state the uncomfortable truth: the child, the drunk, the outsider, the person about to die, the professional with nothing to lose. That licence is a structural resource. Use it once, late.

Third, and this one saves scripts: give the antagonist a case that a reasonable person could hold. If your opposition is stupid or simply cruel, your protagonist wins by default and the audience knows it by page thirty. In Michael Clayton, the character who orders the killings is terrified, over-promoted and out of her depth, which is far more frightening than competence would have been.

Common misconceptions

• Characters must be likeable. They must be comprehensible and active. An audience will follow anyone whose logic they can track and whose want they can feel.
• Backstory deepens a character. Only the fraction that produces present behaviour does. The rest is a document you wrote for yourself, which is fine, as long as it stays out of the script.
• A protagonist must change. A steadfast protagonist under real pressure is a legitimate and ancient shape. What is never legitimate is announcing the change in dialogue.
• The antagonist is the villain. The antagonist is whoever most effectively opposes the want. Frequently it is a person who loves the protagonist.
• Want and need are two names for the same thing. If they were, no choice would ever cost anything, and the third act would have nothing to weigh.

What to carry forward

• Want is external and pursued; need is internal and argued by the film. The gap between them is the engine.
• Characterisation is the surface; character is what a costly choice reveals. Convert value-stating scenes into choices.
• Obstacles sit at inner, personal and social ranges, and the best scripts align all three so one scene advances everything.
• Build from a job with a verb, a contradiction, what they do alone, and what they would never do.
• Arcs may change, hold or fall; the error is announcing the arc rather than dramatising it.
• Every supporting character should want something from the protagonist and refuse them something.
• Give the antagonist a case a reasonable person could hold, or the protagonist wins by default.

Sources

1. Wikipedia contributors. (2024). Michael Clayton (film). Wikipedia. en.wikipedia.org
2. Academy of Motion Picture Arts and Sciences. (n.d.). The Academy. Oscars. oscars.org
3. British Film Institute. (n.d.). Sight and Sound. BFI. bfi.org.uk
4. Egri, Lajos. The Art of Dramatic Writing. Simon and Schuster, 1946. See the chapters on character as the source of premise.
5. Aristotle. Poetics. Translated by Malcolm Heath, Penguin Classics, 1996. On action as the primary element and character as its consequence.

A film that begins in the middle of an argument

The Social Network, written by Aaron Sorkin from Ben Mezrich's book and directed by David Fincher in 2010, opens in a bar with two people already mid-conversation. There is no establishing of who they are, no walking in, no ordering of drinks. The first line lands in the middle of a disagreement about something the audience has not heard the start of, and the scene runs several pages of overlapping, competitive talk. It ends when Erica Albright tells Mark Zuckerberg that he is going to go through life thinking women dislike him for one reason, and that it will actually be another. Then she leaves.

Two things happened. First, the scene entered at the last possible moment and left at the first possible moment, and the audience assembled the missing context without noticing they were doing it. Second, and this is the part beginners skip, the situation at the last line is not the situation at the first. He walks in attached and walks out dumped, and the next twenty minutes of the film are him reacting to it. That is what a scene is for.

The point: A scene is a unit of change. If the situation at the end is the situation at the beginning, you have written an interval, however good the lines are.

What a scene owes you

Take a scene apart and there are four things to find.

A value that moves. Every scene turns on some value that can be positive or negative for someone in it: safety, trust, freedom, hope, respect, love, money, control. The scene opens with that value at one charge and closes with it at the other. Mark a plus and a minus in your margin at the top and bottom of every scene. If both marks are the same, the scene has not happened yet.

Two active parties. At least two people want something in this room, now, and those wants are not compatible. Note the difference between wanting something and wanting something from the other person in the scene.

A turn. The moment the value flips. Usually one line, one action, or one arrival. Find it and you know where the scene is going, which tells you how long the run-up can be.

An exit. The scene ends within a line or two of the turn. Anything after that is the writer being reluctant to leave the room.

A broken scene, printed in full

Here is a scene written badly on purpose. It is the kind of scene every first draft has four of.

INT. SHARED FLAT, KITCHEN - EVENING
JOE, 27, comes in with a bag of shopping. PRIYA, 28, is at the table with a laptop.
PRIYA: Hey. You are late.
JOE: The bus. There were roadworks on Elm Street again.
PRIYA: There always are. Did you get the milk?
JOE: I got the milk. And bread.
PRIYA: Great. I was going to make pasta. Do you want pasta?
JOE: Sure. Actually, listen. I have been meaning to tell you something.
PRIYA: What?
JOE: I am moving out. Rachel asked me to move in with her and I said yes. It would be at the end of next month.
PRIYA: Oh. Wow. Okay.
JOE: I know it is short notice.
PRIYA: No, it is fine. I will find someone. Congratulations, really.
JOE: Thanks. I will still help with the bills until then.
PRIYA: Okay. So, pasta?

Nothing in there is ungrammatical, and every line is plausible. It is still not a scene. Work out why before reading on.

The diagnosis

1. It enters far too early. Six lines of bus, milk and bread precede the only event. The audience will assemble a man coming home with shopping from a single image; they do not need to watch it.
2. Only one person wants anything. Joe wants to deliver news. Priya wants nothing at all until she is told something, at which point she reacts. A scene with one active party is an announcement.
3. The information is handed over free. Joe simply says the thing. Nothing costs him. Compare a version where he needs her to say it is fine, which she withholds.
4. The value does not move for Priya in any way she can act on. She is told, she says congratulations, she offers pasta. Her charge at the top and at the bottom is identical.
5. It leaves late. Three lines after the news, we are still in the kitchen negotiating bills.

Notice that none of these is a dialogue problem. You could rewrite every line with wit and the scene would remain broken, which is why polishing a bad scene never works.

The rewrite

INT. SHARED FLAT, HALLWAY - EVENING
Four flat-packed boxes against the wall. Not new; the tape is old, reused.
PRIYA, 28, coat still on, keys in hand, looking at them.
JOE, 27, in the kitchen doorway. He has been waiting for this for about an hour.
PRIYA: How long have they been in the loft?
JOE: Priya.
PRIYA: It is a fair question. They are your dad's boxes. He gave them to you when you moved in here.
JOE: End of next month.
She puts the keys down on the shelf. Slowly, so it does not look like anything.
PRIYA: Right. Well. That is good. That is good for you.
JOE: You could say it like you meant it.
PRIYA: I could.
A pause. He waits. She does not give it to him.
JOE: I will cover until you find someone.
PRIYA: I know you will.
She goes into her room and shuts the door, quietly, which is worse.
Joe stands in the hall with four boxes.

The scene is now shorter and does more. It enters after the news has been discovered rather than before it is delivered, so the news is not the event; the negotiation over her blessing is. Both parties want something: he wants absolution, she wants him to have to ask for it. The value is approval, and it goes from withheld to explicitly refused, so the charge moves. It exits on an image rather than on logistics. And the information about the bills, which the bad version spent two lines on, survives in a single exchange because it is now a move in an argument rather than an administrative note.

What matters here: Do not fix a scene by improving the lines. Fix it by finding out who wants what from whom, and what they are refusing each other.

Entering late and leaving early, precisely

The advice is universal and usually taught vaguely, so here is a procedure.

To find the entry: write the scene long, then delete from the top, one exchange at a time, until deleting the next one would make the scene incomprehensible. Go back one. That is your entry point. Almost always it is two to four exchanges further in than instinct suggests, and almost always the deleted material was context the audience can infer.

To find the exit: locate the turn. Read forward from it. The moment the scene stops producing new information or new pressure, cut. The last line of a scene is often better as an image than as dialogue, and a scene that ends on an unanswered question pulls the audience into the next one.

There is a related device worth naming: the button, a final line or image that closes the scene with a small snap, often comic or ironic. Buttons are a pleasure and they are also a trap, because a writer who buttons every scene produces a film that feels like a series of sketches. Use them where the tone can carry it and let other scenes end flat, or on an interruption, or by cutting away mid-action, which is frequently the strongest option available.

Beats and tactics inside the scene

Zoom in one more level. A beat is a single exchange of action and reaction: I try something, you respond, and the situation shifts slightly. A scene is a sequence of beats leading to the turn.

What changes between beats is tactic. A character wants one thing throughout the scene, and when a tactic fails they try another. Joe wants absolution: he tries stating the fact flatly, then appealing to fairness, then offering money. Three tactics, one want. If you find a scene going flat in the middle, the usual cause is that a character is repeating the same tactic with different words, which reads as the writer treading water.

A useful working note: list the tactics in the margin as verbs. To confess. To reason. To buy. If two consecutive beats carry the same verb, one of them is redundant.

Three tests, applied fast

The delete test. Take the scene out. If the only loss is information, the scene is a delivery mechanism, and the information should be relocated into a scene that is also doing something else.

The reorder test. Could this scene swap places with the one before it? If yes, neither is causing the other, and you have a sequence of illustrations rather than a story. The fix is usually to make an outcome of one scene the entry condition of the next.

The two-jobs test. Strong scenes do at least two things at once: advance the plot and shift a relationship, or reveal character and plant an object needed later. A scene doing exactly one job is a candidate for merging with its neighbour.

The scenes that seem to break the rule

Some scenes in good films appear to contain no conflict: a meal, a drive, a joke between colleagues, a quiet aftermath. It is worth being accurate about these rather than pretending they do not exist.

Aftermath scenes change a value even when nobody argues, because the audience arrives holding the previous scene. A character alone in a car after a funeral, doing nothing, moves from held-together to not, and the scene is about whether they will let it. Breather scenes buy rhythm, and rhythm is a real function: a film that is tense for ninety unbroken minutes is not tense, it is loud. The rule is not that every scene must contain an argument. The rule is that every scene must be different at the end from how it was at the start, and that difference must matter to somebody.

Common misconceptions

• A scene needs a beginning, middle and end. It needs a turn. Beginnings and ends are largely inferable, which is exactly why you cut them.
• Good dialogue makes a good scene. Dialogue cannot rescue a scene in which nobody wants anything. Diagnose structure before you touch a line.
• Conflict means shouting. Conflict is incompatible wants. The most uncomfortable scenes are usually the politest.
• Every scene should end on a strong line. A film of buttons plays as a set of sketches. Ending flat, or mid-action, is often stronger.
• Quiet scenes have no value change. They change a value the audience is carrying in from the previous scene, which is why they only work in position.

The short version

• A scene is a unit of change: mark the value at the first and last line, and if the charge is the same, nothing has happened.
• Two parties must want incompatible things from each other in the room, now.
• Find the turn, then enter as late as comprehension allows and exit within a line or two of it.
• To find the entry, delete exchanges from the top until the scene breaks, then restore one.
• Beats are exchanges; tactics are the verbs a character switches between while pursuing one want. Repeated verbs mean a flat middle.
• Run the delete, reorder and two-jobs tests on every scene in a draft.
• Quiet and aftermath scenes still move a value, and buttons are a device to ration, not a requirement.

Sources

1. Wikipedia contributors. (2024). The Social Network. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Screenwriting. Wikipedia. en.wikipedia.org
3. British Film Institute. (n.d.). Features and interviews. BFI. bfi.org.uk
4. McKee, Robert. Story: Substance, Structure, Style and the Principles of Screenwriting. ReganBooks, 1997. See the chapters on the scene, the beat and value charge.
5. Yorke, John. Into the Woods: A Five Act Journey into Story. Penguin, 2013.

A diagram drawn in 1979

Syd Field's Screenplay, published by Delacorte Press in 1979, contains a diagram that has probably shaped more scripts than any other single page in the literature. A horizontal bar is divided into three parts. Act One, setup, runs from page one to about thirty. Act Two, confrontation, runs from thirty to about ninety. Act Three, resolution, runs from ninety to about a hundred and twenty. Two vertical arrows drop into the bar near the joins, marked Plot Point One and Plot Point Two.

That diagram did two things at once, one useful and one damaging, and separating them is the whole point of this lesson. The useful thing was to give writers a vocabulary for the shape of dramatic pressure. The damaging thing was to attach page numbers to it, because page numbers turn a description into a schedule, and a schedule can be met without a story ever being told.

Blake Snyder's Save the Cat!, published in 2005, went further in both directions. It set out fifteen named beats with page numbers attached for a hundred and ten page script: opening image, theme stated, set-up, catalyst around page twelve, debate, break into two around twenty-five, B story, fun and games, midpoint around fifty-five, bad guys close in, all is lost around seventy-five, dark night of the soul, break into three, finale, final image. Read as a checklist it produces films that feel prefabricated. Read as a list of questions, it is one of the most efficient diagnostic tools available to a stuck writer.

So we are going to learn the model properly, apply it to a real worked example, and then find out exactly what it is good for.

Key idea: The three-act model is a description of how dramatic pressure tends to accumulate, generalised from films that already worked. It is a diagnostic instrument, and the page numbers are the least important thing about it.

What an act actually is

An act is not a page range and it is not a physical division. It is a unit of action that ends when something happens that changes what the protagonist is doing, so that they cannot continue in the same way. The word comes from theatre, where acts were literally separated by an interval and often by a scene change, but the dramatic unit long predates the interval.

The practical definition to carry: an act ends at a turn after which the protagonist's method or goal is different. If your act break is a big event that leaves your protagonist doing exactly what they were doing before, it is not an act break, it is an incident.

This also explains why the count varies. Some analysts see three acts, some five, some eight sequences. They are looking at the same film and drawing the lines at different levels of resolution, in the way you can describe a journey as two halves, four legs or twenty turnings. The right number is the one that helps you see what is wrong.

The parts, named

Part	What it is	The question it answers
Status quo	The protagonist's ordinary life, shown in action, not described	What is there to lose?
Inciting incident	The event that disturbs the balance and creates the story problem	Why today and not last Tuesday?
Debate or refusal	The protagonist resists, delays, or tries the ordinary solution first	Is the cost real?
First turning point	The commitment. The protagonist chooses a course from which return is expensive	What is the film now about?
Rising action	Attempts, partial successes, new information, escalating opposition	What does trying cost?
Midpoint	A reversal or revelation that changes the terms; often the protagonist stops reacting and starts acting	What did they get wrong at the start?
Second turning point	The worst position: the plan fails, the ally leaves, the truth lands	What is left when the strategy is gone?
Climax	The final confrontation, where the want and the need are finally incompatible	Which one do they choose?
Resolution	The new balance, shown briefly	What did it cost?

Two of these are worth extra attention because they are where drafts die. The inciting incident is often written too late, because the writer is enjoying the world. The midpoint is often missing altogether, which is the single most common cause of a second act that sags, and the fix is almost always to make the protagonist stop reacting to events and start causing them.

Working it: the apiary film, beat by beat

Take the premise from Lesson 2 and build it. A woman inherits her estranged father's failing bee farm with three weeks before the bank takes the land. Want: sell it and clear the debt. Need: find out what her father actually did.

Status quo. KATE, 36, works in commercial property. We meet her ending a phone call badly and then eating standing up. One image tells us she is good at closing things and bad at being in a room.

Inciting incident. Her father dies and the solicitor's letter arrives with the estate: the land, forty hives, and a debt secured against both.

Debate. She tries the ordinary solution, which is to sell it immediately and never go. The buyer wants the land, not the hives, and the sale is conditional on the site surveying clean, which it does not, because of what her father was doing there.

First turning point. To find out why the survey failed, she has to go and stay. She commits to three weeks. Note what makes this an act break: her method changes from selling remotely to being physically present, and everything after it happens on the farm.

Rising action. She cannot open a hive without being stung, and the neighbour, DEREK, who knows how, refuses. She sells her car. She finds her father's hive records, meticulous, going back years, and in the margins another person's handwriting.

Midpoint reversal. She learns that the contamination her father was blamed for came from Derek's land, and that her father took the blame deliberately. Now everything inverts. She came to sell a failure and finds she is standing on somebody's sacrifice. Crucially, she stops reacting: from here she is trying to prove something rather than trying to leave.

Second turning point. The proof, when she finds it, would also destroy Derek, and Derek is the only person who can keep the hives alive past the winter. The bank sets a date. Want and need are now formally opposed: she can have the money or the truth, not both.

Climax. The choice, made in public and at cost.

Resolution. Short. One image, showing what she kept and what she paid.

Notice that the beats were not filled in from a template. Each one came out of the premise, and the model was used only to ask what should be happening at each stage of pressure. That is the correct order of operations, and if you reverse it the film shows.

The upshot: Generate beats from your premise, then use the model to check them. Never generate beats from the model and then look for a premise to hang on them.

What the model is really measuring

Underneath all the vocabulary sits one question: why does the audience keep watching at minute twenty, and again at minute sixty? Every structural device is a proxy for that.

The inciting incident exists because an audience needs a question by minute fifteen. The first turning point exists because a question the protagonist can walk away from is not a question. The midpoint exists because forty minutes of the same kind of pressure is monotonous and attention decays. The second turning point exists because a victory that was never in doubt is not a victory. And the climax must address the specific thing set up at the beginning, because otherwise the audience has spent two hours on the wrong problem.

The oldest version of this is in Aristotle's Poetics, which argues that a plot must have a beginning, a middle and an end, that these must be causally linked rather than merely sequential, and that the events should be probable within the world of the story. Everything since is elaboration.

The three failures the model diagnoses

The slow first act. Symptoms: readers say it takes a while to get going. Cause: the inciting incident is on page twenty-five and the first twenty pages are world-building the writer enjoyed. Fix: move the inciting incident forward and let the world arrive around it. Almost no film has ever been damaged by starting later.

The sagging middle. Symptoms: pages fifty to eighty feel like a series of scenes rather than a story. Cause: usually the protagonist is passive, reacting to whatever the plot brings, and there is no midpoint reversal to change the terms. Fix: give the protagonist a plan at the midpoint, so that the second half of act two is them failing at something they chose rather than absorbing things that happen to them.

The unearned ending. Symptoms: the climax is exciting and feels beside the point. Cause: the climax resolves a problem that was introduced late, rather than the one set up in the first fifteen minutes. Fix: work backwards. Whatever the protagonist most wants to avoid on page ten should be what the climax requires of them.

How the model goes wrong

Three specific dangers, each visible in produced films.

Writing to page numbers produces films that hit marks. You can feel it: at exactly the right moment there is a false victory, and it has no weight because it was scheduled rather than earned.

The template becomes pastiche. When a beat sheet is followed closely, similar films come out, and the audience learns the shape and starts predicting. Prediction is the death of tension.

The model excludes stories it should not. Films built on repetition and routine, on an essayistic argument, on parallel unrelated lives, on a single unbroken situation, or on the deliberate refusal of resolution have all worked at the highest level. If your story does not want a turning point at the end of the first quarter, do not install one. Ask instead what is holding the viewer at minute twenty, and answer that question in whatever form the material offers. The next lesson takes those alternatives seriously.

Common misconceptions

• Three acts means three equal parts. The middle act is conventionally about twice the length of either outer act, and the divisions are dramatic rather than metric.
• An act break is a big event. It is a change in what the protagonist is doing. A large event that leaves the protagonist on the same course is not an act break.
• The model was discovered in ancient drama. Aristotle argued for beginning, middle and end; the numbered three-act paradigm with page counts is a twentieth-century teaching device.
• A beat sheet is a plan. It is a checklist. Beats have to come out of your premise or they arrive as furniture.
• Every film needs a midpoint reversal. Not every film, but almost every film with a sagging middle needs one, which is why it is the first thing to try.

Pulling it together

• An act is a unit of action ending in a turn that changes the protagonist's method or goal; page ranges are conventions, not definitions.
• The named parts are status quo, inciting incident, debate, first turning point, rising action, midpoint, second turning point, climax, resolution.
• Field's 1979 paradigm supplied the vocabulary and the page numbers; Snyder's 2005 beat sheet sharpened both the usefulness and the risk.
• Every device is a proxy for one question: why is the audience still watching at minute twenty and at minute sixty?
• Slow first act means a late inciting incident; a sagging middle usually means a passive protagonist and no midpoint; an unearned ending means the climax answers the wrong setup.
• Generate beats from the premise and check them against the model, never the reverse.
• Films built on routine, essay, parallel lives or refused resolution are legitimate and are the subject of the next lesson.

Sources

1. Wikipedia contributors. (2024). Three-act structure. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Syd Field. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). Save the Cat!. Wikipedia. en.wikipedia.org
4. Field, Syd. Screenplay: The Foundations of Screenwriting. Delacorte Press, 1979.
5. Snyder, Blake. Save the Cat! The Last Book on Screenwriting You'll Ever Need. Michael Wiese Productions, 2005.

A thousand feet of film

A standard reel of 35mm film holds about a thousand feet. Sound film runs at twenty-four frames a second and there are sixteen frames to a foot, which works out at ninety feet a minute, so a full reel runs about eleven minutes before somebody has to change it. For decades that was the physical unit of cinema: films were shot, assembled, shipped and projected in reels, and audiences in the early period watched them one reel at a time.

Writers adapted. If the projectionist is going to interrupt you every eleven minutes, each stretch has to be satisfying on its own, with its own small question raised and answered. That habit outlived the technology, and it became the eight-sequence approach, taught by Frank Daniel at the University of Southern California and at Columbia and set out in Paul Gulino's Screenwriting: The Sequence Approach in 2004. A two-hour film, eight sequences, each running twelve to fifteen minutes, each a small film with a beginning, a rise and a resolution that opens the next.

Why this matters: The three-act model tells you where the big turns go. It tells you nothing about what to do for the forty minutes in between. The sequence approach fills exactly that gap.

The eight-sequence approach, worked

A conventional distribution across a feature, with the three-act landmarks marked so you can see how the models sit on top of one another.

Sequence	Its own question	Three-act landmark
1	Who is this and what is their world?	Status quo, inciting incident
2	What is the problem and will they take it on?	Debate, first turning point
3	What is the first plan and how does it fail?	Early act two
4	What is learned, and what does it cost?	Approach to midpoint
5	What changes now that the terms are different?	Midpoint reversal
6	What happens when the new plan meets real opposition?	Late act two
7	What is left when everything is lost?	Second turning point
8	How does it end and what did it cost?	Climax and resolution

Take sequence three for the apiary film. Its question: what is Kate's first plan and how does it fail? Her plan is to sell the land without ever learning anything about it. The sequence opens with the buyer's surveyor arriving, runs through her attempts to get the paperwork signed off, includes the small humiliation of not being able to open a hive in front of a stranger, and ends when the survey fails and the sale collapses. That is twelve to fifteen minutes with its own arc, and it hands the next sequence a new question: if she cannot sell it remotely, what will she have to do?

The practical value is enormous, because a stuck writer usually knows the big turns and is lost in between. Asking what question this twelve-minute stretch answers is a smaller, more answerable question than asking what should happen in act two.

John Yorke and the five-act shape

John Yorke, who ran drama production at the BBC and worked across many long-running British series, published Into the Woods in 2013 arguing that the underlying shape of story is five acts rather than three, and, more interestingly, that structure is not an arbitrary convention but an expression of how change happens.

His five movements run roughly: the protagonist in a state of lack; the disruption and the reluctant entry into a new world; the acquisition of a false or partial solution, peaking at a midpoint moment of truth where they see themselves clearly; the consequences of that knowledge, which strip away the false solution; and the final test in which the change is either made or refused.

Two of his claims are worth carrying whatever model you use. First, the midpoint is not an event but a recognition: the moment the protagonist glimpses what they actually are, which is why films sag there when the writer inserts a car chase instead. Second, the shape is fractal. The same movement of desire, complication, recognition and consequence appears in a scene, in a sequence, in an act and in the whole film, which is why a good scene feels like a small version of the story it is inside.

Non-linear construction

Three films, three different jobs.

Akira Kurosawa's Rashomon, released in 1950 and adapted from two stories by Ryunosuke Akutagawa, tells one incident four times from four incompatible points of view. The non-linearity is the subject: the film is about the unreliability of testimony, and a chronological version would be a different and much lesser film.

Quentin Tarantino's Pulp Fiction, from 1994, with story credit shared with Roger Avary, reorders three interlocking crime stories so that a character killed in one section is alive in the next, and so that the film ends in a diner scene that chronologically sits near the middle. The reordering buys tonal control and a comic irony that would be unavailable in sequence, and it lets the film end on grace rather than on death.

Christopher Nolan's Memento, from 2000 and based on his brother Jonathan Nolan's story, runs two strands: black and white sequences moving forward, colour sequences moving backward, meeting at the end. The structure puts the audience inside the protagonist's condition, since like him they arrive at each scene without knowing what preceded it.

From those three cases, one rule and one procedure. The rule: non-linearity must buy something the chronological version cannot deliver, and the something is almost always a manipulation of what the audience knows and when. If your reordering only makes the film harder to follow, it is decoration.

The procedure: write the story out in chronological order first, in an outline, then choose the presentation order deliberately, marking for each scene what the audience knows at that point that the characters do not, and the reverse. That marking sheet is the actual craft document of a non-linear film. Working without it is how writers end up with a mystery whose solution the audience worked out forty minutes early.

Remember: Chronology is a resource, not a default. Spend it only where withholding or reordering changes what a scene means.

Ensemble and mosaic

Robert Altman's Nashville, from 1975, follows around two dozen principal characters across five days in one city, with the strands touching, glancing off each other and finally converging. Steven Soderbergh's Traffic, from 2000, with a screenplay by Stephen Gaghan adapted from the British television series Traffik, runs three largely separate stories about the drug trade, distinguished on screen by colour treatment, which converge thematically rather than mechanically. Films of this kind are sometimes grouped under the label hyperlink cinema.

Writing one presents problems the single-protagonist models do not address.

The connection problem. Why are these stories in the same film? There are three acceptable answers: they share a cause, they share a place, or they are three answers to one question. The weakest answer is coincidence, which is why the coincidence-driven mosaic has aged worse than the other kinds.

The heat problem. An audience does not weight strands equally. One will always be the one they are waiting to return to. You can either accept that and structure the film around it, giving the hot strand the ending, or deliberately starve it, which is a real strategy and an unforgiving one.

The rhythm problem. Cutting between strands has its own timing. The usual working rule is to leave a strand on a question and return to it later than the audience wants, and never to visit a strand only for housekeeping. If you return to a story and nothing has changed since we left it, cut the visit.

The length problem. Each strand needs a minimum viable size to be felt at all. In practice this means fewer strands than you first plan. Three strands well built beat six sketched.

Four more shapes worth knowing

• Repetition and routine. The film shows the same activity many times, and meaning accumulates from small differences. Structure comes from variation rather than escalation, and the ending is usually a break in the pattern.
• Single situation, near real time. One place, one continuous problem. Structure comes from the arrival of new information or new people, and the writer's discipline is to ration entrances.
• The frame narrative. A story told inside another. It buys a narrator whose reliability can be questioned, and it costs momentum every time you return to the frame, so the frame must earn its returns.
• The picaresque. A protagonist travels through episodes that do not cause one another. Held together by voice and by a cumulative portrait rather than by causation, which is why it lives or dies on character.

Choosing a shape

If the material is...	Consider	Because
One person pursuing one goal against opposition	Three-act, checked with eight sequences	The models were generalised from exactly this case
One person who must see themselves differently	Yorke's five movements	It puts recognition, not incident, at the centre
About testimony, memory or complicity	Non-linear	Order controls what the audience knows, which is the subject
About a system rather than a person	Mosaic, three strands maximum	No single protagonist can see the whole of a system
About endurance, labour or grief over time	Repetition and variation	Escalation would falsify a subject that does not escalate

Common misconceptions

• The alternatives are for art films. The eight-sequence approach is taught in mainstream film schools and describes a great deal of commercial cinema.
• Non-linear means complicated. It means ordered for effect. Some of the clearest films ever made are non-linear, and some incomprehensible ones are strictly chronological.
• An ensemble film has no protagonist. It usually has a hottest strand that functions as one, and pretending otherwise is how ensembles lose their ending.
• Choosing a non-three-act model frees you from structure. It replaces one set of obligations with another, generally stricter one. Repetition films in particular are unforgiving about variation.
• The models compete. They sit on top of one another. Most working writers hold three acts, eight sequences and Yorke's recognition beat in mind at the same time.

Looking back

• The eight-sequence approach descends from eleven-minute reels and fills the gap the three-act model leaves in the middle.
• Each sequence poses and answers its own question, which is a smaller and more answerable problem than what happens in act two.
• Yorke's five movements place a recognition at the midpoint and treat the same shape as fractal across scene, sequence, act and film.
• Non-linear order must buy a change in what the audience knows and when; write it chronologically first, then choose the order with a knowledge sheet.
• Mosaic films need a real answer to why these stories share a film, an acknowledged hot strand, and no more than about three strands.
• Repetition, single-situation, frame and picaresque shapes each impose their own discipline rather than removing structure.
• Choose the model that fits the material, and expect to hold two or three of them at once.

Sources

1. Wikipedia contributors. (2024). Nonlinear narrative. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Hyperlink cinema. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). John Yorke. Wikipedia. en.wikipedia.org
4. Wikipedia contributors. (2024). Rashomon. Wikipedia. en.wikipedia.org
5. Gulino, Paul Joseph. Screenwriting: The Sequence Approach. Continuum, 2004.
6. Yorke, John. Into the Woods: A Five Act Journey into Story. Penguin, 2013.

A document somebody pays for

Under the Writers Guild of America's Minimum Basic Agreement, a screenwriting deal is normally broken into steps, and each step is a separate delivery with its own payment: a story or treatment, then a first draft screenplay, then a rewrite, then a polish. This matters to you even if you never sign such a contract, because it tells you something the how-to books usually leave out. A treatment is not merely a planning aid a writer keeps in a drawer. It is a document the industry treats as a stage of the work, reads carefully, and forms opinions from.

So there are two families of planning document, and confusing them causes real damage. Private documents are for finding out whether a film exists. Delivered documents are for making somebody else see the film you have found. They contain roughly the same information and they are written completely differently.

The point: Work out who a planning document is for before you write a word of it. The same content written for the wrong reader is wasted.

Four documents, and what each is for

Document	Length	Reader	Job
Logline	One or two sentences	Anyone, including strangers	Make somebody want to hear more
Beat sheet	Half a page to two pages	You, usually	Hold the shape of the whole film in one glance
Step outline	Three to ten pages, one entry per scene	You, and sometimes a collaborator	Test whether the scenes cause one another
Treatment	Five to twenty-five pages of prose	Producers, financiers, directors, competitions	Make a reader feel the film in the right order and the right tone

Two warnings about the last row. A treatment written like a synopsis, all "then he goes to see her, and she tells him," is dead on the page and will kill the project in the room. And treatment length is decided by purpose: a producer who asked for something short means it, and sending twenty-five pages when six were requested reads as an inability to judge a room.

The step outline, worked

A step outline gives one entry per scene, and a good entry is not a description of what happens. It names who wants what, what stops them, and what has changed by the end. Two or three sentences. Here are the first ten steps of the apiary film, written properly.

1. KATE closes a property deal by phone in a stairwell, efficiently and coldly, then eats a sandwich standing up. She is good at ending things.
2. A solicitor's office. Kate wants the estate turned into money. The solicitor explains the debt is secured against the land and the hives, which cannot be separated. She leaves owning something she does not want.
3. Kate wants to sell without going. A phone call to a developer: he is interested, conditional on a clean site survey. She agrees to a date three weeks out. The clock starts.
4. Kate arrives at the farm to let the surveyor in. Forty hives, unattended. She cannot get near them. She waits in the car until he is finished, which tells us everything about her.
5. DEREK, the neighbour, walks over to say the hives will die within a month untended. Kate wants him to keep them alive until the sale. He refuses, without explaining, and goes.
6. The survey comes back failed: contamination on the south field, historic, unresolved. The developer withdraws unless it is cleared. Kate now needs to know something about her father to sell her father's land.
7. Kate wants documents. She goes through the house and finds no paperwork but forty years of hive records, weather, yields, mite counts, in her father's hand, and margin notes in a second hand she does not recognise.
8. Kate wants the environmental file and goes to the council office. It exists, it is closed, and getting it opened needs the signature of the other party named in it. The other party is Derek.
9. Kate asks Derek to sign. He wants her to open a hive first, which is either a condition or a joke. She does it badly and is stung across the wrist. He signs nothing.
10. Kate sells her car. Small, quiet, and the first thing in the film she has done that costs her.

Read that back and notice three things you can only see in an outline. There is a want in every step. Step six is where the want and the need first touch, which is early enough. And steps four and five could be merged, which is exactly the kind of decision that is cheap to make now and expensive to make on page forty.

Cards, and why physical ones still win

The card method is old and still the fastest way to restructure. One scene per index card. On each card, four things: the location, who wants what, the obstacle, and the change. Lay them out in rows on a table or a board, one row per act or per sequence.

What the physical version gives you, and the software version mostly does not, is a whole-film view at a glance plus the ability to move ten cards in five seconds. Colour is worth using: one colour per strand or per point of view, and a board with all the cards of one colour clustered in one third of the film has told you about a pacing problem you would not have found by reading.

Two working habits. Write the cards fast and badly the first time, because the value is in the reordering. And keep a separate pile for cards you love and cannot place, which stops you from bending a structure to accommodate a scene you are fond of. Some of those cards belong in a different film, and admitting it early is a kindness to yourself.

Writing the treatment

A treatment is prose, present tense, third person, telling the story in the order the film tells it. It is not a report about a film. It is the experience of the film, compressed.

Concretely, that means several things. Write in scenes, not in summary: name the place, put people in it, let something happen. Use dialogue sparingly, and only when a specific line is the point; three or four lines in a ten-page treatment, each chosen because it carries the film's voice. Keep the tone of the finished film. A comedy treatment should be funny on the page. A thriller treatment should withhold. If your treatment reads as neutral, you have written a summary and the reader will assume the film is neutral too.

Two structural decisions that catch people out. First, treatments normally reveal the ending. They are not marketing to an audience, they are informing a decision-maker, and a treatment that coyly declines to say how the film ends reads as a writer who does not know. Second, the first page has to work harder than the rest, because a reader who is not held by page one reads the remaining nine with reduced attention, and no amount of quality on page seven can undo that.

The reliable test: read it aloud to somebody who has not heard the idea, without stopping to explain. Anywhere you find yourself adding a clarification out loud is a place where the document is failing, and you should write the clarification in and delete something else.

Worth holding on to: A treatment should read as a story told well, in the tone of the film, with the ending included.

Two ways outlining goes wrong

The list of events. Every entry says what happens and none says who wanted what. Such an outline looks complete and teaches you nothing, because the question it was supposed to answer, do these scenes cause one another, cannot be asked of it. The diagnosis is easy: if you can read your outline without ever encountering the word wants, or a synonym, rewrite it.

The outline so complete the script is dead. Some writers plan until every line is decided, then discover that writing the script is transcription, and transcription is joyless work that shows. The practical rule that most working writers converge on: outline until you are bored of outlining, then start writing. The boredom is a real signal. It means the discovery has moved from the plan into the pages.

There is a third, subtler failure: outlining as avoidance. If you are on your fifth outline of the same film and no pages exist, the outline has stopped being a tool and become a place to hide. The fix is a deadline and a page count, not a better outline.

When to abandon the plan

You will be forty pages in when a scene wants to go somewhere the outline does not allow. Sometimes that is the material teaching you something, and sometimes it is the writer avoiding a difficult scene by taking an interesting exit.

The way to tell them apart is cheap. Follow the impulse for ten pages, then stop and ask what the film now needs after those pages. If you can re-outline from there and the story is stronger, you were right and the outline was wrong. If you cannot, you have found out for the price of ten pages, which is nothing.

What you must not do is keep writing forward for sixty pages with an outline you know is dead. That is how first drafts become unfinishable, and it is the single most common way promising scripts are abandoned.

Common misconceptions

• Outlining kills spontaneity. Outlining relocates spontaneity from structure, where mistakes cost months, into scenes, where they cost an afternoon.
• A treatment is a summary. A summary reports; a treatment performs. If yours reads neutral, the reader will assume the film is neutral.
• Treatments should hide the ending to create interest. A treatment is read by somebody deciding whether to spend money. Withholding the ending reads as not having one.
• Longer treatments are more thorough. Length is set by who asked and why. Sending twenty-five pages when six were requested is itself information about you.
• The outline should be followed to the end. Follow a promising deviation for ten pages and then re-outline. Abandoning the plan silently for sixty pages is how drafts die.

What you now know

• Planning documents split into private ones, for finding the film, and delivered ones, for making somebody else see it.
• Logline, beat sheet, step outline and treatment differ by length, reader and job, and the industry treats the treatment as a paid step.
• Every step outline entry should name a want, an obstacle and a change, not just an event.
• Cards give a whole-film view and fast reordering; colour by strand exposes pacing problems reading cannot find.
• Treatments are present tense, in scenes, in the tone of the film, with sparing dialogue and the ending included.
• Outline until you are bored, then write. Repeated outlining with no pages is avoidance.
• When the material fights the plan, follow it for ten pages and re-outline from there.

Sources

1. Writers Guild of America West. (n.d.). Writers Guild of America West. WGAW. wga.org
2. Wikipedia contributors. (2024). Treatment (film). Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). Spec script. Wikipedia. en.wikipedia.org
4. Field, Syd. Screenplay: The Foundations of Screenwriting. Delacorte Press, 1979. On the card method and structuring by index card.
5. Gulino, Paul Joseph. Screenwriting: The Sequence Approach. Continuum, 2004.

Three lengths of silence

Harold Pinter, who wrote for the stage and also wrote the screenplays for The Servant in 1963, The Go-Between in 1971 and The French Lieutenant's Woman in 1981, and who was awarded the Nobel Prize in Literature in 2005, wrote three different stops into his texts and treated them as three different instructions. Three dots marked a hesitation inside a thought. A marked Pause was longer, and meant something had happened between the characters. A marked Silence was longer again, and usually meant somebody had lost.

The point is not that you should imitate his notation. It is that Pinter was writing the pressure, not the words. In his scenes people talk about the weather and the tea and a mislaid coat while an entire negotiation about power runs underneath, and the gaps are where the negotiation is visible. If you take one idea from this lesson, take that one: dialogue is what people do to each other with speech, and the words are the least of it.

The core of it: A line of dialogue is an action. If you cannot say what a line is doing to the other person, it is not doing anything.

Dialogue is an action, not a transfer

Beginners write dialogue as information moving between characters. Working writers write dialogue as one person trying to change another person's state, and the information travels as a side effect.

The practical version is the verb test. Take any speech and write an active verb in the margin describing what the speaker is doing to the listener. To warn. To flatter. To stall. To humiliate. To recruit. To apologise without apologising. If the only verb you can find is "to explain" or "to tell," you have a delivery mechanism rather than a line.

Run this down a page and two problems become visible immediately. Lines with no verb, which should be cut. And consecutive lines with the same verb, which means a character is repeating a tactic, and repeated tactics read as the writer treading water even when the wording varies.

Subtext, defined so you can use it

Subtext is the gap between what a character says and what they want. That definition is operational: it tells you how to make it. Give a character something they need from the other person and a reason they cannot ask for it directly, and subtext appears without any effort on your part. The reasons people cannot ask directly are ordinary and numerous: pride, politeness, fear, the presence of a third person, professional role, the fact that asking would concede something.

On the nose is the opposite: a character states their emotional condition, their motive, or the theme of the film out loud. "I have never felt safe since my father left." Real people almost never do this, and when they do, it is a dramatic event rather than a piece of routine dialogue.

A useful exercise is the two-column pass. Take a page of your dialogue and, beside each line, write what the character actually wants at that moment. Where the two columns say the same thing, you have on-the-nose writing. Where they differ, you have subtext. A page in which they differ on every line is probably evasive to the point of vagueness. A page in which they match on every line is a lecture. Most good pages differ on most lines and match, once, at the moment somebody breaks.

The same scene twice

A woman wants her sister to take their mother for the summer. Version one:

ANNA: I need you to take Mum for the summer. I have had her for two years and I am exhausted and it is not fair.
BEV: I know it is not fair, but I have my own problems, and you have always been better with her than me. I feel guilty about it all the time.
ANNA: Guilt is not the same as helping.

Every line is honest, articulate, and dead. Nobody hides anything, so nothing is at stake in the speaking. Version two:

ANNA: She has started putting the milk in the cupboard.
BEV: Everyone does that.
ANNA: Twice a day?
Bev refills the kettle that does not need refilling.
BEV: Is she eating?
ANNA: I make her eat.
BEV: You should get someone in. Just for a couple of days a week. I can look into it.
ANNA: With what?
BEV: I said I would look into it.
ANNA: The spare room at yours gets the sun all afternoon.
Silence. Bev stares at the kettle.
BEV: It is not really a spare room.

Nobody has said take her or I cannot. The whole negotiation happened, the request was made and refused, and both women are now in a worse position than when they started. Note that the second version is not more clever than the first; it is more constrained. Anna cannot ask outright because asking outright would let Bev refuse outright, and once refused she has nothing left. That constraint is what generated the milk, the kettle and the spare room.

What matters here: Subtext is not a decorative style. It is what happens when a character has a good reason not to say the thing.

How to hear a bad line

Seven tests, in the order that finds problems fastest.

1. Read it aloud. Not silently, actually aloud, at speaking pace. Half of all bad lines are audible immediately, because your mouth will refuse them.
2. Cover the character names. Read the dialogue column alone. If you cannot tell who is speaking, you have one voice wearing several names.
3. Ask who the line is for. If the answer is the audience rather than the other character, it is exposition wearing a costume, and the next lesson is about fixing that.
4. Check whether it answers the question. People deflect, change the subject, answer a different question or answer a question with a question. A page where every line directly answers the previous one is a questionnaire.
5. Ask what it is doing. The verb test. No verb, no line.
6. Look at the length. Real speech is uneven. If every speech is two to three lines, the page will have a droning quality. Vary between one word and, rarely, a genuine long speech, which should be reserved for a character who has earned the floor.
7. Ask whether an actor could do it without the line. If a gesture, a pause or a look would carry the same content, cut the line. This test alone will remove ten percent of any first draft's dialogue and improve all of it.

Making people sound different

Distinct voices are not achieved by giving one character a catchphrase. Five variables do almost all the work.

Variable	What to decide
Vocabulary band	Which register do they reach for under pressure: technical, bureaucratic, religious, sporting, domestic?
Sentence length	Do they run on, or stop early? Does that change when they are lying?
Completion	Do they finish their sentences? Do they finish other people's?
What they never say	The words a character avoids are more characterising than the ones they use. A man who never says the word wife.
Register under stress	Everyone shifts. Some go formal, some go crude, some go quiet, some go into jargon they can hide inside.

Fill that table in for your three main characters before you write a word of a scene between them and the name-blind test will start passing on its own.

Rhythm, interruption and overlap

Scripts mark an interruption by cutting the interrupted line off, conventionally with a dash, and starting the next speaker immediately. Simultaneous speech is written as dual dialogue, side by side in two columns, and should be used sparingly because it is hard to read.

More useful than either device is the deliberate mismatch of rhythm. Give one character short lines and the other long ones and the scene acquires a shape before you have decided anything about content. Reverse it at the moment the power changes hands. In the sisters scene above, Anna has the short lines throughout, until the last exchange, where Bev gets the last and shortest line and wins.

Silence is an instrument you own. In format you have several ways to write it: an action line describing what someone does instead of speaking, a beat, a plain Silence on its own line. Prefer the first, because it gives the actor something to play rather than an instruction to be still.

Accent, dialect, profanity

Do not spell accents phonetically. Written-out dialect is hard to read, ages badly, and frequently reads as contempt for the speaker. Convey regional and class speech through word choice, idiom, grammar and rhythm instead: which words a person reaches for, how they order a sentence, whether they use the full form or the contraction. A single well-chosen local noun does more than a page of dropped letters.

Profanity is a rhythm and a register decision, not a realism decision. Characters who swear constantly quickly stop registering as swearing at all, which wastes an effect you may want later. Decide what each character's ceiling is, and let one of them break it once.

Common misconceptions

• Good dialogue is realistic dialogue. Transcribed real speech is unusable: repetitive, circular and full of nothing. Dialogue is designed speech that gives the impression of the real thing.
• Subtext means being vague. It means a character has a specific want and a specific reason not to say it. Vagueness has no want underneath it and reads as evasion by the writer.
• Characters should sound different because of catchphrases. Vocabulary band, sentence length, completion, avoidance and stress behaviour do the work; verbal tics are the cheapest and least durable option.
• A long speech is a sign of weak writing. A long speech is a resource. It works when a character has earned the floor and everybody in the scene has a reason to let them keep it.
• Phonetic spelling captures an accent. It slows the reader and often patronises the character. Word choice and syntax carry accent far better.

Summing up

• Every line is an action; mark it with a verb, and cut lines whose only verb is to explain.
• Subtext is the gap between what is said and what is wanted, produced by giving a character a reason they cannot ask directly.
• On-the-nose dialogue states motive, feeling or theme out loud, and should be reserved for the moment somebody breaks.
• Seven tests find bad lines: read aloud, cover the names, ask who it is for, check for deflection, apply the verb test, vary length, and ask whether an actor could do it silently.
• Distinct voices come from vocabulary band, sentence length, completion, avoidance and behaviour under stress.
• Rhythm mismatch between characters shapes a scene before content does, and reversing it marks a shift in power.
• Convey accent through word choice and syntax rather than phonetic spelling, and ration profanity so it keeps its force.

Sources

1. The Nobel Prize. (2005). The Nobel Prize in Literature 2005: Harold Pinter. Nobel Prize Outreach. nobelprize.org
2. Wikipedia contributors. (2024). Harold Pinter. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). Subtext. Wikipedia. en.wikipedia.org
4. Mamet, David. On Directing Film. Viking, 1991. On telling the story in images and distrusting explanatory speech.
5. McKee, Robert. Dialogue: The Art of Verbal Action for Page, Stage, and Screen. Twelve, 2016.

A bomb under a table

In the long series of conversations Alfred Hitchcock had with Francois Truffaut, published in 1966 and in English in 1967, Hitchcock set out a comparison he had used for years. Four people sit at a table talking about baseball. A bomb under the table goes off. The audience gets fifteen seconds of shock. Now run the same scene, but first show the audience the bomb and the clock. Nothing else changes. The same people have the same dull conversation about baseball, and it is unbearable, because every word about baseball is now a word not spent getting out of the room.

That is an argument about exposition, though it is usually filed under suspense. What changed between the two versions was not the plot and not the dialogue. It was who knew what, and when. Exposition is not the problem of getting information into an audience. It is the problem of deciding what they know, what the characters know, and how the difference generates pressure.

Key idea: Handle information as a resource with timing, not as a debt to be discharged as early as possible.

What exposition actually is

Exposition is information from before the story, or from outside the scene, that an audience needs in order to understand what is happening. It comes in three kinds, and they have different difficulty levels.

Kind	Example	Difficulty
World rules	How the magic works, what the law says, what a night shift involves	Easiest: show it operating once and the audience generalises
Backstory	The affair nine years ago, the bankruptcy, the war	Hardest: nobody in the room has a reason to say it
Present logistics	Where they are going, who is meeting whom, what the deadline is	Medium: easy to state and easy to make dull

Before you write a single explanatory line, run the audit. List every fact you think the audience needs. Beside each, write the latest scene at which they could learn it and still follow the film. Then move each fact to that scene. You will find that a third of your list was never needed at all, and that half of the rest can arrive later than you assumed, which is where the tension was hiding.

The audience surrogate problem

The standard bad solution is a character who exists to be told things: the new recruit on their first day, the visiting journalist, the amnesiac, the child. Somebody explains the world to them, and the audience listens in.

This device has a long history. The nineteenth-century well-made play often opened with two servants tidying a room and telling each other things they both already knew, so that the audience could be brought up to date. The modern version begins "As you know, Bob," and every writer has written one.

Why does it fail? Not because it is old, and not because newcomers are unrealistic. It fails because in such a scene nobody wants anything. The explainer is happy to explain, the listener is happy to listen, nothing is refused, nothing costs, and the scene has no charge at either end. It is the delivery mechanism from Lesson 4 in its purest form.

Which tells you how to fix it rather than merely to avoid it. A newcomer is a perfectly good device as long as the newcomer wants something and the explainer does not want to give it. The recruit who needs to know the procedure because they have already broken it. The journalist who is not being told, and has to work it out from what the press officer refuses to say. Those are scenes. The difference is not the device; it is the presence of opposition.

Six deliveries that work

1. Put it in conflict. Somebody wants the information and somebody else does not want to give it. The audience learns the fact and learns two characters at the same time, and the scene has a charge.
2. Give it to somebody who has it wrong. A character states the backstory inaccurately and is corrected, or worse, is not corrected. Now the fact arrives with an argument attached, and the correction is a dramatic act.
3. Attach it to an object or an action. A hospital wristband still on somebody's arm three weeks later. A second toothbrush. A car seat in the back of a van driven by a man who says he has no children. Objects deliver backstory silently and the audience feels clever, which is a pleasure worth giving them.
4. Withhold and let them infer. Audiences are much faster than writers expect. If two people avoid a subject with sufficient consistency, the audience will construct the subject. The risk is real ambiguity, so pay it off later.
5. Deliver it late, when it changes meaning. The same fact told on page eighty recolours everything before it. This is the strongest and most economical option and the one beginners use least.
6. Make it cost the speaker. If saying it out loud is a humiliation, a confession, or a betrayal, the exposition becomes the event of the scene rather than a pause in it.

A worked repair

Here is exposition written the wrong way. Two detectives in a car.

RAY: You know Halloran has been on this squad twenty-two years, right? He was the one who caught the Delaney case in 2011, and after his wife died he requested nights and he has been on nights ever since.
SAM: I heard about his wife. That is rough.
RAY: He is a good detective. He just does not talk to anybody.

Everything is true and nothing is happening. Now the same facts, repaired using three of the six techniques.

SAM: Halloran took my file off the desk.
RAY: Then it is his file.
SAM: I have been on this squad eight months.
RAY: He has been on it since before you could drive.
Sam looks out of the window.
SAM: Everyone keeps telling me he is good. Nobody tells me anything else.
RAY: What else do you want?
SAM: Somebody in the canteen said he asked for nights. Who asks for nights?
Ray does not answer. He turns the radio on, which is an answer.
SAM: Right.

The audience now knows Halloran is senior, respected, works nights by choice, and that there is something everybody knows and nobody says. It does not know about the wife, and it does not need to yet; that fact is now worth something, and it should be spent much later, ideally by Halloran, at cost. Meanwhile the scene has two people wanting incompatible things: Sam wants to be told, Ray wants not to be the one who tells him.

The upshot: Exposition delivered by somebody who does not want to deliver it stops being exposition and becomes a scene.

Voice over, honestly assessed

Voice over has a bad reputation among teachers of screenwriting, and the reputation is deserved for one specific misuse and undeserved otherwise.

The misuse is narration that tells you what you are already watching. If the image shows a man walking into a house and the voice says he walked into the house that day not knowing what he would find, the film is doing the same job twice and the audience is being managed.

The legitimate uses are real and several. Voice over can supply a second layer that contradicts or complicates the image, which is where irony lives. It can establish a narrator whose reliability is itself in question. It can compress time efficiently, moving a story over months in twenty seconds. And it can create a relationship between the teller and the audience that the drama alone cannot.

Billy Wilder's Sunset Boulevard, released in 1950 and written by Wilder with Charles Brackett and D. M. Marshman Jr., is the standard demonstration. The film opens on a body in a swimming pool and the narrator is the dead man, telling you how he ended up there. Everything he says is coloured by the fact that he does not survive the film, so the narration is not redundant with the image: it is a second position on it.

The working test: cover the voice over and watch the scene. If nothing is lost, cut the narration. If what is lost is a fact, find another way to deliver the fact. If what is lost is a relationship, an irony or a tension between what is said and what is shown, keep it.

How much can you withhold?

New writers explain too much in the first ten pages because they are afraid of confusing a reader. The genuine requirement is much smaller than it feels. In the opening an audience needs, roughly: who to follow, what kind of film this is, and a question they want answered. They do not need names, jobs, histories, or the rules of the world, all of which can arrive across the next thirty minutes as they become relevant.

The counterweight is that confusion is different from mystery. Mystery is knowing what you do not know. Confusion is not knowing what you are supposed to be attending to. An audience will happily wait an hour for an answer to a question they can hold in their head, and will not wait five minutes if they cannot tell what the question is.

So the discipline is: withhold answers freely, and be extremely clear about the question.

Common misconceptions

• Exposition must be got out of the way early. Most of it should be moved as late as it can go, because a fact that arrives late recolours everything before it.
• A newcomer character is always a mistake. A newcomer who wants something, from someone who does not want to give it, is a scene. The fault is passivity, not the device.
• Audiences need to be told or they will not follow. They infer very fast, and they enjoy inferring. What they will not tolerate is not knowing which question the film is asking.
• Voice over is lazy. Narration duplicating the image is lazy. Narration in tension with the image is a second layer that no other tool provides.
• Flashbacks solve backstory. A flashback stops the present story and has to earn the interruption. Prefer an object, an error corrected, or a confession that costs.

The takeaway

• Exposition is a timing problem: what the audience knows, what the characters know, and the pressure the gap creates.
• Audit every fact, find the latest scene at which it could arrive, and move it there. A third of the list will turn out to be unnecessary.
• The surrogate fails because nobody in the scene wants anything, not because newcomers are unrealistic.
• Six working deliveries: conflict, a character who has it wrong, objects, inference, late arrival, and a cost to the speaker.
• Voice over earns its place when it adds a layer the image cannot: irony, unreliability, compression or a relationship with the teller.
• In the opening an audience needs a person to follow, a sense of the kind of film, and a question. Nothing else is urgent.
• Withhold answers freely; never be unclear about the question.

Sources

1. Wikipedia contributors. (2024). Hitchcock/Truffaut. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Exposition (narrative). Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). Sunset Boulevard (film). Wikipedia. en.wikipedia.org
4. Wikipedia contributors. (2024). Voice-over. Wikipedia. en.wikipedia.org
5. Truffaut, Francois, with Helen G. Scott. Hitchcock. Simon and Schuster, 1967.

Two words that sold a film

Alien, released in 1979 and directed by Ridley Scott from a screenplay by Dan O'Bannon and a story by O'Bannon and Ronald Shusett, was famously pitched with a phrase of three words: Jaws in space. Jaws had opened four years earlier and everybody in the room had seen it.

Consider what that phrase actually transmitted. A small group in a confined place they cannot leave. A creature that is mostly unseen. Members of the group killed one at a time, in an order the audience will try to predict. Authority that will not act until it is too late. A survivor who is competent rather than heroic. A final confrontation in the creature's element rather than ours.

Not one of those was stated. They were all understood, because both people in the conversation had a shared model of what that kind of film promises. That shared model is what genre is, and it is the reason genre is worth studying even if you intend to write something that has no name yet.

The point: Genre is not a shelf you are put on. It is a set of expectations your audience arrives holding, which you did not create and cannot ignore.

Promises, not ingredients

The unhelpful way to think about genre is as a list of ingredients: a western has horses, a horror has a monster. That model fails immediately, because films with all the ingredients routinely feel wrong and films with none of them feel exactly right.

The useful model is a contract of promises about the experience. A horror film promises dread and the violation of safety. A thriller promises that a competent person is in escalating danger and that you will be ahead of them sometimes and behind them at others. A romantic comedy promises that two people who should be together will be kept apart by something that is their own fault, and that they will end up together, or that the film will pay heavily for refusing. A detective story promises that the truth is discoverable and that you will be given a fair chance at it.

Three consequences follow. Genre is separable from setting: a western can be set in a hospital and a heist film in a monastery. Genre is separable from tone: two horror films can be funny and solemn respectively while making the same promise. And genre is decided by the audience's expectation, not by your intention, which is why the diagnosis section below matters more than it sounds.

Obligatory scenes

Every genre carries scenes the audience will feel cheated without. They are not a formula: how you deliver them is entirely open, and the pleasure is often in an unexpected delivery. But skipping them is a decision with a price, and you should know you are paying it.

Genre	Core promise	Obligatory scenes	Most common failure
Horror	Safety will be violated and dread will be sustained	The first violation; the rules being learned; someone not believed; the false ending	Showing the threat too early and too completely
Thriller	Escalating jeopardy for someone competent	The discovery; the failed appeal to authority; the isolation; the reversal of hunter and hunted	A protagonist who is not good at anything
Detective	The truth is discoverable and you get a fair chance	The scene of the crime; the false solution; the interview that changes everything; the accounting	Withholding a clue the detective had, which breaks the contract
Romantic comedy	Two people kept apart by their own faults will meet properly	The meeting; the reason it cannot work; the near miss; the decision to try	An obstacle that is external, so nobody has to change
Heist	Competence as pleasure, plus the risk of betrayal	Assembling the team; the plan explained; the plan going wrong; the part of the plan we were not shown	A plan the audience could not follow, so the reversal means nothing
Coming of age	A person will cross a line and not be able to go back	The world before; the older figure who is not what they seemed; the humiliation; the small act of adulthood	Nostalgia in place of cost
Courtroom	Argument as combat, with a verdict	The impossible instruction; the witness who breaks; the evidentiary problem; the verdict and its aftermath	Legal procedure invented so freely that the stakes evaporate

Read down that failure column, because it is more useful than the rest of the table. Almost every one is the same underlying error: the writer delivered the surface of the genre while dropping the thing it exists to produce.

Which genre did you actually write?

Writers regularly intend one genre and produce another, and the discovery usually arrives from a reader in the least welcome terms. You can find it yourself with two questions.

First: at any given point in the script, what does the audience want to happen next? Not what is interesting, what do they want. If, in your intended thriller, the answer on page fifty is that they want the two leads to talk honestly to each other, you have written a relationship drama with a plot attached, and you should either honour that or restore the jeopardy.

Second: what would make them feel cheated? Write the list. If your list does not match the obligatory scenes of the genre you thought you were writing, trust the list.

The reason this matters commercially as well as artistically is that films are financed, cast, marketed and reviewed against genre expectation. A film sold as one thing and delivering another is punished by audiences who wanted the first thing and never reached by the audience who wanted the second.

Why this matters: Your genre is what your audience is hoping for, not what you told them at the start.

Subversion, and the price of admission

Subverting a genre is one of the most reliable ways to make something feel new, and it has one hard condition: you must have delivered the genre well enough that the audience believes you could have gone on delivering it. A promise you never made cannot be broken; it can only be absent.

The Coen brothers' No Country for Old Men, adapted in 2007 from Cormac McCarthy's novel, is the clearest teaching example. The film is built as a pursuit thriller and executes the form with total competence for most of its length, and then declines to show the confrontation the entire structure has been promising, cutting instead to the aftermath, and ends on a man describing a dream. That works because two hours of impeccable thriller earned the refusal. The same ending on a film that had been sloppy about tension would read as an inability to write the scene.

So the practical rule is a sequence: deliver, deliver, deliver, withhold. And know what you are trading. A subverted ending buys resonance and costs satisfaction, and both of those are real currencies. Decide which one your film needs and do not pretend the exchange is free.

Mixing

Most films are mixtures, and the mixing question that actually matters is which promise governs the ending. A romantic comedy with a crime plot ends on the relationship. A crime film with a romance ends on the crime. When a film gets this wrong it feels as if it stopped rather than ended, and the usual cause is a writer resolving the plot they found more interesting to write rather than the one the audience was watching.

Tone is a separate contract and audiences police it more strictly than plot. A film may move between comedy and horror, but the rules of movement have to be established early. If your first fifteen minutes are entirely comic and the first violent death arrives at minute forty, the audience does not read it as bold, they read it as a mistake. Establish the range in the first act and you can then use all of it.

What genre does to structure

Genre is not only a matter of content; it changes the machine.

• Pace of information. A mystery can withhold for a long time because withholding is the pleasure. A comedy cannot, because confusion is not funny.
• Position of the protagonist. Thrillers and horror usually keep the audience slightly ahead of the protagonist to create dread; detective stories keep them slightly behind to create curiosity.
• Length of the setup. Genres with a strong shared model can start faster, because the audience arrives already knowing the rules. A film in an invented world spends longer.
• Where the midpoint sits. In horror the midpoint is often the moment the rules are understood, after which the film becomes a problem to be solved rather than a mystery to be survived.

Common misconceptions

• Genre is commercial and therefore lesser. Genre is a communication system between a film and its audience. Some of the most formally adventurous films ever made are genre films.
• Obligatory scenes are a formula. They are a set of moments an audience will feel cheated without. Their content and their handling are entirely yours.
• Subversion is a shortcut to originality. Only a delivered promise can be broken. Withholding a scene you never earned reads as an inability to write it.
• A film can be genre-neutral. Audiences classify within minutes whether you cooperate or not. Refusing to decide means somebody else decides for you, usually the marketing department.
• Mixing genres means splitting the difference. One promise has to govern the ending. Films that refuse to choose feel as though they stopped rather than finished.

Where this leaves us

• Genre is a contract of promises about the experience, held by the audience before the film starts.
• Promises, not ingredients: genre is separable from setting and from tone, and it is decided by expectation rather than intention.
• Each genre carries obligatory scenes; skipping one is a decision with a price you should know you are paying.
• Diagnose your real genre by asking what the audience wants to happen and what would make them feel cheated.
• Subversion requires prior delivery, and trades satisfaction for resonance.
• In a mixture, one promise must govern the ending, and tonal range must be established in the first act.
• Genre changes the machinery: pace of information, whether the audience is ahead or behind, setup length and the function of the midpoint.

Sources

1. Wikipedia contributors. (2024). Alien (film). Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Film genre. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). No Country for Old Men. Wikipedia. en.wikipedia.org
4. British Film Institute. (n.d.). Features and interviews. BFI. bfi.org.uk
5. Altman, Rick. Film/Genre. BFI Publishing, 1999.

A book with no plot, adapted anyway

Susan Orlean published The Orchid Thief in 1998. It grew out of a magazine assignment about a Florida orchid poacher called John Laroche, and it is a very good book, but consider what it is: a writer circling an obsession, digressing into the history of orchid collecting, admitting she does not feel about anything the way Laroche feels about flowers. Nobody in it wants a specific thing, is blocked from getting it, and runs out of time.

The book was optioned, and Charlie Kaufman was hired to adapt it. What he produced, filmed by Spike Jonze as Adaptation in 2002, is a film about a screenwriter called Charlie Kaufman failing to adapt The Orchid Thief, equipped with a twin brother named Donald who is cheerfully writing the kind of thriller Charlie despises. The screenplay was credited to Charlie Kaufman and Donald Kaufman, and both names appeared on the Academy Award nomination for Best Adapted Screenplay, which makes Donald the only invented person ever nominated for one. Chris Cooper won Best Supporting Actor for playing Laroche.

Do not copy the solution. Copy the first move. Before Kaufman wrote anything he worked out what the book really was, and concluded that the film he had been hired to write did not exist inside it. That diagnosis is where every adaptation starts, and skipping it is why so many adaptations are accurate and dead.

So what?: An adaptation begins with a diagnosis of the source, not a decision about how faithful to be.

What each form does without effort

Prose and screen are not two languages saying the same things. They are good at different things, and the mismatch is the entire job.

Capacity	Prose does it free	Screen does it free
Interior state	She had never forgiven him. Direct, unmediated, unarguable.	Nothing. It must be converted into behaviour, object or speech.
Time	Twelve years passed. Four words.	Nothing free. A card, a cut, a changed face, a season.
Digression	A chapter on the history of orchid smuggling, then back.	Nothing. The screen cannot stop and think without stopping.
Point of view	Can move between heads at will, or lock into one absolutely.	Defaults to omniscient watching. Locking a film to one consciousness takes constant work.
The physical world	Costs a paragraph per room and slows the page.	A whole room, era, class and weather arrive in one frame at no cost.
A face reacting	Costs a sentence and risks telling the reader what to feel.	One cut. The most efficient thing cinema owns.
Simultaneous behaviour	Must be serialised: first this, then that.	Six people doing six things at once, all legible.

Read that table as a conversion job with two directions. Everything in the prose column either gets a physical form or gets cut. Everything in the screen column is free money you should be spending: an adaptation that keeps four pages of a character's thoughts and none of the room they are sitting in has the exchange backwards.

The arithmetic is subtraction

A hundred and ten page screenplay contains, generously, twenty thousand words, and a great deal of the page is white space by design. A mid-length novel runs four or five times that, and the difference is not padding. Which gives the single most useful rule in adaptation: a feature film holds roughly the quantity of story in a novella or a long short story. Adapting a novel is therefore mostly amputation. Adapting a short story is mostly expansion, which is why short stories are the friendliest source material a beginner can find and novels are the least forgiving.

Two worked cases. James Ellroy's 1990 novel L.A. Confidential is long, densely plotted and carries a large cast. Brian Helgeland and Curtis Hanson's screenplay cut whole plot lines and characters out of it and narrowed the film onto three policemen, Bud White, Ed Exley and Jack Vincennes. It won the Academy Award for Best Adapted Screenplay for 1997. Nobody who loves the film considers it a reduction, because what was cut was not the engine.

Second: Peter Benchley's 1974 novel Jaws contains an affair between the oceanographer Hooper and Brody's wife, and a strand of organised crime pressing on the town's finances. The 1975 film, written by Benchley and Carl Gottlieb, drops both. That looks like simplification and is actually a structural repair. Once Hooper is not sleeping with Brody's wife, the three men on the boat can become something the audience wants to survive together, and the final act works. Cutting the affair did not lose material. It bought the third act.

Bottom line: Cutting is not damage control in adaptation. It is the primary creative act, and the best adaptations cut more than their authors first thought possible.

The load-bearing wall test

Here is the procedure. It takes an afternoon and it is worth more than any theory of adaptation.

1. Write down six to ten things a reader who loves this source would say it is. Effects, not plot points: the dread, the voice, the sister, the last line, the fact that you never see the creature.
2. Mark each one structural or decorative. Structural means that removing it makes other things fall down. Decorative means that removing it costs pleasure and nothing else.
3. Cut every decorative element without argument, however good it is. Your affection for it is not evidence.
4. For each structural element, name the physical means by which a camera and a microphone will deliver it. Be specific: not the loneliness, but the empty second chair kept at the table.
5. Whatever survives step four with no physical means attached is your real adaptation problem. There is usually exactly one, and the film succeeds or fails on how you solve it.

Run it on a story you know. Edgar Allan Poe's The Tell-Tale Heart, first published in 1843, is about eight pages long, and the list comes out something like this: the narrator's insistence that he is sane; the old man's clouded eye; the seven nights of the door opened an inch at a time; the killing; the floorboards; the police drinking tea in the room; the heartbeat only the narrator can hear; the confession.

Nearly all of that is filmable at once. The door, the eye, the boards, the police, the confession: a competent student could shoot them next week. But work through step two honestly and you find that the engine is none of those. It is that the whole story is a voice protesting its own sanity in a manner that demonstrates the opposite, and a camera pointed at a man on a staircase records a man on a staircase. That is the one thing with no physical means attached, and every filmed version of the story is really an answer to it.

Four honest answers to the interiority problem

When you find your one unfilmable element, you have four moves. They are not ranked by quality; they are ranked by how often they are available.

1. Externalise it into behaviour. Find the action that only a person in that state would perform. Lesson 1 did this with a handshake given at the wrong temperature. This is the best answer whenever it exists, because the audience does the work and therefore believes the result.
2. Give the thought a listener. A thought is inert; an argument is a scene. Poe's narrator, placed in an interview room with a duty solicitor who plainly thinks he is unwell, turns the entire story into present-tense pressure, with the murder as what he claims happened. The voice survives, and it now costs him something to speak.
3. Give it an object. A thing kept, hidden, repaired or refused carries interior states across a whole film without a word. Kate's refusal to leave the car while the surveyor walks the field, in the film we have been building since Module 1, does more than a page of her thinking about her father.
4. Use voice over, structurally and on purpose. Lesson 9 was hard on voice over used to patch a hole. This is the other case. John Huston's last film, The Dead, released in 1987 from a screenplay by Tony Huston, adapts the closing story of Joyce's Dubliners, published in 1914, and ends with Gabriel's interior passage spoken over the film. The story's ending is a thought, and the film decides to say so rather than pretend otherwise. That is a choice made at the level of the whole film, not a bandage applied in the edit.

In short: Name your unfilmable element, then choose one of these four in writing, before you draft. Adaptations that fail rarely chose badly. They mostly failed to notice there was a choice.

Fidelity is the wrong question

Reviewers ask whether an adaptation is faithful. It is not a question you can act on, because it has no failure condition you could test on page forty. Replace it with one you can: does the film do to its audience what the source did to its readers?

By that standard a rearrangement can be faithful and a transcription can be a betrayal. Emma Thompson's screenplay for Sense and Sensibility, filmed in 1995, moves Austen's material around freely, compresses, invents business and ages a character, and it is faithful in the only sense that matters, because the experience survives. It won the Academy Award for Best Adapted Screenplay, making Thompson the only person to have won Academy Awards for both acting and writing.

The same test frees you in the other direction. You are allowed to adapt a bad book. Plenty of great films come from ordinary material that contained one strong situation, and there is no obligation to preserve the parts that were not working on the page either.

Two techniques do most of the practical cutting. Amalgamation folds two or three characters into one, which is almost always available and almost always improves a film, because a character with two functions is more interesting than two characters with one each. Relocation moves a scene's content into a setting that is already earning its place, so that an expository conversation happens during the only funeral you can afford to shoot.

Rights, briefly, because beginners get this wrong

You cannot adapt copyrighted material and sell the result. In practice a producer takes an option: a payment for the exclusive right, for a fixed period, to try to set the film up, with an agreed purchase price if it happens. Public domain material carries no such problem, which is why Austen, Poe and Dubliners are adapted constantly. Real events are not copyrightable, but life rights are bought from living subjects anyway, for cooperation, access and legal comfort.

None of this forbids you from adapting a novel you love as a writing exercise. It only means you have produced a sample rather than a saleable script, which is a legitimate use of six months as long as you know which of the two you did.

Common misconceptions

• Adapting is easier than original writing because the story already exists. You inherit the source's problems along with its strengths, and you cannot fix a broken second act by invention alone without being accused of infidelity.
• Faithful means good. The most beloved passage in a book is frequently the least filmable thing in it, and reproducing it exactly is how adaptations die politely.
• Novels are the natural source and short stories are too thin. Usually the reverse. A short story gives you a situation and room to build, which is invention with a spine already in place.
• You have to include the famous scene. Only if it is load-bearing. If it is decorative, its fame is a reason to look at it harder, not a reason to keep it.
• Voice over means you failed. Voice over used to explain what an audience could have seen is a failure. Voice over chosen as the film's relationship to a consciousness is a strategy, and it has to be decided early and held.

What to remember

• Diagnose the source first: what it actually is, and whether a camera can reach its engine.
• Prose and screen are free at different things. Convert or cut what prose gets free; spend what the screen gets free.
• A feature holds about a novella. Novels require amputation, short stories require expansion.
• The load-bearing wall test separates structural from decorative and isolates your single real problem.
• Four answers to interiority: externalise, give it a listener, give it an object, or commit to voice over as structure.
• Judge by what the film does to its audience, not by how closely it tracks the page.
• Amalgamate characters and relocate scenes: the two cuts that almost always pay.

Sources

1. Wikipedia contributors. (2024). Adaptation (film). Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Film adaptation. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). L.A. Confidential (film). Wikipedia. en.wikipedia.org
4. British Film Institute. (n.d.). Features and interviews. BFI. bfi.org.uk
5. Seger, Linda. The Art of Adaptation: Turning Fact and Fiction into Film. Henry Holt, 1992.

Eighteen minutes that bought a feature

Damien Chazelle had a feature screenplay about a young drummer and a violent teacher that nobody would finance. So he took one section of it, shot it as a short film of about eighteen minutes, and sent it to Sundance, where it premiered in 2013 and won the festival's jury prize for short film. The feature was financed on the strength of it. Whiplash premiered at Sundance the following year and won the Grand Jury Prize; J. K. Simmons won the Academy Award for Best Supporting Actor for the part.

Notice what Chazelle did not do. He did not compress the feature into eighteen minutes. He took one confrontation out of it and made that the whole film. The short has a beginning, a middle and an end of its own, and it is not a trailer or a summary. That distinction is the entire subject of this lesson, and getting it wrong is why most short scripts fail before anyone reads page two.

Remember: A short film is not a small feature. It is a different form, and its governing law is that it holds one change, completely, rather than several changes partially.

Three reasons to write one, and what each one implies

Be honest about which of these you are doing, because they pull the writing in different directions.

Purpose	What it is for	What it forces you to write
A complete work	The story is this size and no other. Nothing precedes or follows it.	An ending that is genuinely an ending, and material with no need for a wider world.
A proof of concept	Demonstrating that a feature idea works, in order to get the feature made.	The most compelling single sequence in the feature, shot at feature quality, standing alone.
A calling card	Showing what you can do so that someone gives you the next job.	Whatever you can execute to the highest standard, which usually means small, controlled and performance-led.

Writers get into trouble by attempting all three. A proof of concept that also tries to be complete ends up explaining the feature's world. A calling card that also tries to prove a concept reaches for scale it cannot afford and looks thin. Choose one before you outline.

What length actually means

The Academy of Motion Picture Arts and Sciences defines a short film for its awards as a film of forty minutes or less, including credits. That is the outer boundary, not a target. In practice the useful band is much narrower, for a reason that has nothing to do with art: festival programmers assemble shorts into blocks of about ninety minutes, so a twenty-five minute film has to be substantially better than a nine minute film to earn its place, because it eats the space of two or three others.

So write short. Seven to twelve pages is the sweet spot for a first film, and the discipline of the page count is not a limitation on the story, it is a filter on which stories you should be telling. If your idea needs twenty-two minutes to work, ask whether it needs them because it is genuinely that shape, or because you have not yet found where it starts.

The single change

A feature can carry a want, a need, a subplot, a reversal at the midpoint and a changed protagonist. A short can carry one thing changing, and the change should be observable rather than internal, because you have no time to establish an interior baseline and then move it.

The shape that works most reliably has four beats and no fat.

1. Situation. Who, where, and what the ordinary rule is. One page, ideally delivered while something else is happening.
2. Pressure. The rule is tested. Two to four pages, escalating.
3. Turn. A single decision or discovery. One page, usually one line or one action.
4. Consequence. What the person does in the world that they would not have done before. One to two pages, then stop.

Two structural rules follow. First, no subplots, at all. Second, start as late as you possibly can: a short is usually the third act of something, and the material that would be acts one and two in a feature is what the audience infers. Two or three characters, one or two locations, and one night or one afternoon.

A worked short, built from a premise you have already met

Lesson 2 tested this idea for dramatisability: a night nurse discovers a colleague stealing morphine. As a feature it would need an investigation, an institution, a cost, an ending. As a short it needs nine pages, and it stops at the moment of change.

1. Page 1. A controlled drugs register on a shelf. Two signatures are required per entry. NAOMI, night shift, signs the second line without reading the first, because she is being spoken to at the same time. The rule of the world, delivered as an action, not as information.
2. Page 2. A ward round. A patient in genuine pain asks when the next dose is. Naomi checks the chart. The times on it do not match what the patient is telling her.
3. Page 3. She counts the ampoules in the cupboard. One short. She counts again, slower. Still one short. Nothing is said aloud for a page and a half.
4. Page 4. She looks at the register entry for the missing one. Her own signature is on it.
5. Pages 5 and 6. MARTIN arrives with two coffees, tired, kind, the person who covered her shift in March. He asks her to sign the next entry now, so he does not have to come and find her later. He holds out the pen.
6. Page 7. She signs. The turn is not the signing. The turn is that after he goes, she takes out her phone and photographs the page.
7. Pages 8 and 9. The patient's bell. She gives the injection properly, and afterwards, alone, she puts the empty ampoule in her pocket instead of the sharps tray. Cut.

Count what that film contains. Two characters. One location and its corridor. One night. No exteriors, no crowds, no vehicles. One change, which is fully observable: a woman who was covered by the system becomes a woman collecting evidence, and remains complicit while doing it. And it ends on an action rather than on a revelation, so it does not need a scene explaining what happens next.

The core of it: Find the single moment in your material when a person becomes someone slightly different, put it on page seven, and build only what is needed to make it land.

Four ways short scripts fail

Read a hundred submitted shorts and the same four problems account for most of them.

1. The twist dependency. The film exists to deliver a final revelation: he was dead, it was a dream, the killer was the narrator. Because everything before the twist is only setup, nothing before the twist is interesting, and the audience spends nine minutes waiting to be told the point. Test: if your last thirty seconds were removed, would there still be a film? If not, you have a joke with a long approach.
2. Beginning at the beginning. The character wakes, dresses, travels, arrives, and the actual material starts on page five of a nine page film. Apply the entering-late discipline from Lesson 4 twice as hard here, because you have a fifth of the running time to work with.
3. The mood piece. Beautiful images, weather, a person walking, no want. Atmosphere is a multiplier of story, not a substitute for it. A film with nobody wanting anything is a sequence of good shots that an audience watches politely.
4. Writing what cannot be shot. The most common single cause of unmade shorts. Crowds, night exteriors, moving vehicles, rain on demand, animals, small children, period costume and stunts are all expensive or slow or both, and a nine page script that needs three of them will not be made by a first-time crew.

Production reality is a writing tool

Every line you write is a purchase order. A scene heading that reads EXT. MOTORWAY BRIDGE, NIGHT, RAIN commits somebody to a permit, a location manager, a rain machine, lighting a bridge and a night shoot. That is not a reason to write timidly. It is a reason to write deliberately.

The productive method is to invert the usual order. Write down what you can actually get: a flat, a car park after hours, a launderette, a church hall, two good actors, two days, one afternoon of a friend's van. Then build a story whose pressure comes from those things. A film built out of exactly what you can get is more likely to be good than a diluted version of a film you cannot afford, and constraint is generative rather than merely restrictive: the corridor and the drug cupboard above are not compromises, they are why that short is tense.

One practical note on dialogue. A short is usually made with the actors you can persuade rather than the actors you would cast, so write dialogue that survives an ordinary performance. Lines that need a great actor to be bearable are a risk you are transferring to somebody who has not agreed to take it.

Endings, which is what people remember

Shorts are judged disproportionately on their last twenty seconds, because that is the part of the experience the viewer is left holding. Three endings that tend to work: an action that costs the character something; an image that reframes what you have already seen without contradicting it; and a decision made and not explained. Two that tend to fail: the reveal that retrospectively empties the film, and the slow fade on a face doing nothing, which is what a script writes when it has not decided how it ends.

Worth holding on to: Write your last shot before you write page one. If you cannot name it, you have a situation rather than a short film.

Common misconceptions

• A short is a feature in miniature. Compressing three acts into nine minutes produces a synopsis with actors in it. One change, told completely, beats three changes told partially.
• Shorts are for beginners. Established writers and directors use them as proofs of concept, and the form has its own festival and awards ecology that runs independently of features.
• A strong twist makes a short. Twist dependency is the single most common failure in submitted shorts, because everything before the twist becomes setup and setup is not interesting to watch.
• Constraints hurt the writing. The limits decide which stories you tell, and stories chosen for what you can actually shoot are the ones that get finished and seen.
• Length is a badge of ambition. A twenty-five minute film competes for the space of two or three others in a programmer's block, so length is a cost you have to justify.

Recap

• A short holds one observable change, built as situation, pressure, turn, consequence, with no subplots.
• Decide first whether you are writing a complete work, a proof of concept or a calling card; attempting all three produces none of them.
• Forty minutes is the Academy's outer limit for a short; seven to twelve pages is where a first film should live.
• Start as late as possible. A short is usually the third act of something, and the audience infers the rest.
• The four failures: twist dependency, beginning at the beginning, the mood piece, and writing what cannot be shot.
• Write from what you can get. Constraint is material, and the limits are why the tense films are tense.
• Name your last shot before page one, and end on an action or a decision rather than an explanation.

Sources

1. Wikipedia contributors. (2024). Short film. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Academy Award for Best Live Action Short Film. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). Whiplash (2014 film). Wikipedia. en.wikipedia.org
4. Sundance Institute. (n.d.). Sundance Institute. sundance.org
5. Cooper, Pat, and Ken Dancyger. Writing the Short Film. 3rd ed., Focal Press, 2005.

The committee that exists because rewriting is normal

When a film arrives with three names on its writing credit, somebody had to decide the order, the wording and the ampersands. The Writers Guild of America settles it by arbitration: the participating writers submit their drafts, three Guild members read the material anonymously, and they determine who contributed enough to be credited and how the credit reads. An entire institutional apparatus exists for this.

Think about what that implies. A trade body does not build an anonymous three-person adjudication system for something that happens occasionally. It builds one because rewriting, including rewriting by other hands, is the normal condition of a screenplay rather than an embarrassment that befalls weak ones. Your first draft is not a script that needs tidying. It is the raw material from which a script is made, and the making is this lesson.

Bottom line: The first draft's job is to exist. Every good screenplay you have ever seen was bad at some point and was made good afterwards, deliberately, in passes.

Why polishing is not rewriting

Here is the trap that costs writers years. You open the draft at page one, start improving lines, feel productive, reach page thirty, lose momentum, and start again at page one next month. Six months later the first thirty pages are beautifully written and the film still does not work, because none of the problems were on the line level.

Lesson 4 made the point at scene scale: you can rewrite every line of a broken scene with wit and the scene remains broken. The same is true of a script. So the method is not to improve the draft, it is to run a sequence of passes, each answering one question over the whole script, with everything else forbidden during that pass.

Pass	The one question	Forbidden during it
1. The cold read	What is this film about now, and where exactly did I stop caring?	Changing a single word
2. Structure	Does each act break turn the story, is the midpoint a real reversal, can any sequence be removed whole?	Touching dialogue
3. Character	What does each person want in each scene, and who wants nothing anywhere?	Adding new scenes
4. Scene	Does this scene change a value, and can it enter later and leave earlier?	Fixing lines
5. Dialogue	Can this line be cut, halved, or given to the other character?	Adding information
6. Polish	Do the action lines read at speed, and is the format clean?	Rethinking anything

The order is not arbitrary and it is not negotiable. Work upward from the largest unit, because every hour spent perfecting a line inside a scene you will later cut is an hour spent making the cut harder to face. And the prohibition column matters as much as the question column: you cannot judge whether a scene earns its place while you are enjoying its best joke.

The cold read, and the only marks you may make

Print it if you can. Read the whole thing in one sitting, at reading speed, without stopping. You are not editing; you are impersonating a stranger, which is difficult and gets easier if you leave the draft alone for two weeks first.

You may make exactly three marks. A wavy line down the margin where your attention went. A circle where you were confused about who, where or why. A tick where you sat up. No notes, no fixes, no promises to yourself.

The next day, work only from the marks, and read them correctly. A wavy line at page fifty-five rarely means page fifty-five is badly written; it usually means that whatever was supposed to be at stake stopped being at stake somewhere around page forty. Circles cluster where you cut something in an earlier draft and did not replace what it was carrying. Ticks are the most useful marks of all, and the most ignored: they tell you what your film is actually good at, which tells you what the rest of it should be serving.

The deletion test at three scales

Ask of every unit: if this were gone entirely, what breaks? Then be strict about which answers count.

• Line. If cutting the line loses nothing but a beat of personality, cut it and see whether the personality survives elsewhere. It usually does, and the scene gets faster.
• Scene. The scene survives only if removing it breaks a causal chain or removes a value change. The answer that does not count is: we would not know that Kate's father kept records. Information can arrive almost anywhere, which is precisely why scenes that exist to deliver it are the ones to cut.
• Sequence. The hardest and most valuable. Take five to fifteen pages at a time and ask whether the film survives their removal. Most bloated second acts contain one full sequence that is doing what a neighbouring sequence already does, better.

There is a version of this you can run in an afternoon. List every scene on one line each with its page count, print the list, and read it as a document. Scenes that repeat a function are visible on a list in a way they are never visible inside a draft.

A revision pass, worked in full

Here is a scene from the first draft of the film this course has been building since Module 1. Kate has to get the environmental file opened, and the file needs Derek's signature.

INT. DEREK'S KITCHEN - DAY
Derek is making tea. Kate stands by the door with a folder.
KATE: Thanks for seeing me.
DEREK: Sit down if you want.
KATE: I am fine. I went to the council office about the survey. The contamination on the south field.
DEREK: I know about the survey.
KATE: They said the file is closed and they cannot open it without a signature from the other party named in it. That is you.
DEREK: It is.
KATE: I need the file open or the sale falls through. If the sale falls through the bank takes the land, and then the hives go anyway. So I am asking you to sign it.
DEREK: I am not going to do that.
KATE: Why not?
DEREK: Because I am not.
KATE: That is not an answer. My father is dead. Whatever happened between you, it was twenty years ago.
DEREK: You do not know what happened.
KATE: Then tell me.
DEREK: No. Do you want this tea or not?
KATE: No. Thank you.
She leaves. Derek stands with two cups.

Every line is playable and the scene is doing its plot work. It is also the scene a first draft always produces, and here is what the passes find.

Pass 2, structure. The scene survives the deletion test, but only just: it delivers a refusal that the story needs. Note for later that its function is one line long.

Pass 3, character. Kate wants a signature. Derek wants nothing at all. He is a wall for her to hit, which is the commonest character failure in a middle act. Give him a want and the scene becomes an encounter instead of a request.

Pass 4, scene. It enters four lines early, on greetings. It also has no physical activity, so both actors have nothing to do but speak, which is what makes a scene feel like a table read. And the value change is thin: she wanted the signature at the top and does not have it at the bottom, so the only movement is a small loss.

Pass 5, dialogue. Kate states her entire situation in one speech, including the bank, the sale and the hives, all of which the audience already knows. Derek says he is not going to do it, then says it again. And the last exchange explains its own subtext out loud: then tell me, no.

Now the rewrite. Notice that the plot outcome is identical. He still does not sign.

EXT. DEREK'S APIARY - DAY
Twelve hives in a row, in far better condition than Kate's father's. Derek works a frame, unhurried, no gloves. Kate stands upwind holding a smoker he has handed her.
DEREK: Squeeze it. Not like that. Twice, then stop.
She squeezes it twice. Smoke. He lifts a frame out and reads it the way other people read a letter.
KATE: The council will not open the file without your name on a form.
DEREK: I know what the council needs.
He holds the frame out toward her. She does not take it.
DEREK: Forty hives your father had. You have not opened one.
KATE: I am selling them.
DEREK: You are selling them.
He puts the frame back. Closes the hive. Takes the smoker out of her hand, which takes a moment, because she does not let go of it immediately.
DEREK: Come back when you have opened one.
KATE: Is that a condition or are you being clever?
He walks the row, checking the next lid, not looking at her.
DEREK: It is a condition.

Here is the accounting, change by change, so you can copy the reasoning rather than the scene.

Change	Pass	What it bought
Moved from kitchen to apiary	4	Physical activity for both actors, and a location that argues Derek's case without a word: his hives are thriving.
Gave Derek a want	3	He wants her to touch the bees. Two active parties, so the scene is a negotiation rather than a request.
Cut the greeting and the folder	4	Four lines and the sense of a scene beginning at its beginning.
Cut Kate's speech about the bank and the sale	5	Six lines of information the audience already had. What remains is one clause of plot.
Turned the refusal into a condition	2	A refusal ends a strand; a condition starts one. The scene now generates the next three scenes instead of closing a door.
Handed her the smoker in the first line	4	The scene's real subject, whether she will touch the hives, is running underneath from the top.
Cut then tell me, no	5	The withheld history stays withheld and is now more interesting, because nobody has drawn attention to the withholding.
Ended on him walking the row	4	Leaving early. The old version ended on two cups of tea, which is a writer waiting for a better exit.

So what?: The rewrite kept the plot and changed almost everything else. That ratio is normal. When a scene fails it is rarely the outcome that is wrong; it is the want, the activity, the entry point and the amount said out loud.

Reading notes from other people

You will be given notes, by friends, readers, producers and eventually by people paying you. The single most useful principle: a note is reliable evidence about where a problem is and unreliable evidence about what it is. Somebody who says the sister should have more scenes has usually detected that the sister is inert when present, which more scenes would make worse.

So translate before you act. Two questions extract more than any general request for feedback. Where did you get bored, exactly, page number if you can? And at page thirty, what did you think was going to happen? The second is the diagnostic one, because it tells you what expectations your script is actually setting, as opposed to the ones you meant to set.

Then decide by pass. If three readers circled the same twenty pages, that is structure and belongs to pass 2. If one reader disliked a character everyone else liked, that is taste and belongs to nobody. If you find yourself defending a scene out loud with information that is not in the script, the scene is not working and the defence is the proof.

When it is a page-one rewrite

Sometimes the passes are not the answer. If the cold read tells you that the film is about something other than what you have been writing, or that the protagonist is the wrong person, no sequence of scene-level repairs will get you there, and you are looking at a page-one rewrite: the same idea, drafted again from a blank page, usually keeping ten or fifteen scenes you can lift wholesale.

It is less painful than it sounds, and it is faster than the alternative, because the second version of a scene you already understand takes an afternoon rather than a week. What makes it feel unbearable is the sunk cost, which is not a real argument. The material you cut is not wasted; it is what taught you the film.

Knowing when to stop

There is a point where passes start producing lateral changes rather than improvements, where you swap one adequate line for a different adequate line and cannot tell which is better a day later. That is the signal to stop. A draft with one genuine strength and one visible flaw does better with readers than a smooth draft with neither, because the strength is what makes someone want to work on it and the flaw is what makes them feel useful.

In short: Rewrite in passes, from the largest unit down, using the cold read and the deletion test, translating notes into diagnoses. Then stop while the film still has an edge on it.

Common misconceptions

• Rewriting means improving the writing. Most of what a rewrite does is structural and characterological. The line-level pass comes fifth for a reason.
• A scene that delivers necessary information has earned its place. Information can arrive almost anywhere, so that is the weakest possible justification for a scene and usually the sign of one that should be cut.
• Notes should be executed as given. Notes locate problems and misname them. Execute the diagnosis, not the prescription, and expect the two to differ.
• A page-one rewrite means the draft was wasted. The draft is what taught you the film, and the second version of an understood scene takes hours rather than days.
• You should rewrite until nothing is wrong. Past a certain point you are making lateral swaps, and sanding away the flaws often takes the strengths with them.

What to carry forward

• Rewriting is the normal condition of a screenplay, not a remedy for a weak first draft.
• Run passes, one question each, from largest unit to smallest, with everything else forbidden during the pass.
• The cold read allows three marks only: boredom, confusion, and the places you sat up.
• Apply the deletion test at line, scene and sequence scale, and refuse the excuse that a scene delivers information.
• When a scene fails, the outcome is usually right and the want, the activity, the entry point and the said-out-loud portion are wrong.
• Turning a refusal into a condition converts a closed door into three more scenes.
• Notes are accurate about location and unreliable about cause. Ask where they were bored and what they expected to happen.
• Stop when the changes go lateral, and ship a draft with an edge rather than one sanded flat.

Sources

1. Writers Guild of America West. (n.d.). Writers Guild of America West. WGAW. wga.org
2. Wikipedia contributors. (2024). Writers Guild of America. Wikipedia. en.wikipedia.org
3. British Film Institute. (n.d.). Features and interviews. BFI. bfi.org.uk
4. Goldman, William. Adventures in the Screen Trade. Warner Books, 1983.
5. McKee, Robert. Story: Substance, Structure, Style, and the Principles of Screenwriting. ReganBooks, 1997.

The document about your script that you will never see

Send a screenplay to a production company and, if it is read at all, it produces a second document. Coverage is usually two pages. At the top sit the title, the writer, the form, the genre, the setting and a one sentence logline written by the reader rather than by you. Then a synopsis of half a page to a page. Then the comments, which is the only part anyone reads twice. Then a grid, typically rating premise, structure, characterisation and dialogue as excellent, good, fair or poor. Then a verdict with three possible values: pass, consider, recommend.

Two facts about that document should change how you write. The first is that the reader is often a freelancer paid per script, working through several over a weekend, and their professional reputation depends on not recommending things that waste an executive's Saturday. The second is that the executive reads the coverage, not the script, unless the coverage tells them to read the script. Your screenplay is competing to be summarised well by a tired stranger.

Recommend is rare. Most readers issue few of them in a year, because a recommend is a bet on their own judgement. Consider is the realistic good outcome, and it means the executive may read it themselves or ask for the writer's other material. Pass is not an insult; it is the default state of the universe.

The core of it: Your script's first job is to survive a professional summariser. Everything you cannot make survive a two page precis is a decoration you are hoping somebody will notice.

How readers actually read

Nobody who reads scripts for money reads them the way you read your own. The realistic pattern is that a verdict forms early and the rest of the read confirms or disturbs it. That is not laziness; it is what happens to anyone reading their fourth script of the day, and it produces three practical consequences.

• The first ten pages decide the frame. Not because page sixty does not matter, but because a reader who has decided at page ten that this is competent will read page sixty generously, and a reader who has decided the opposite will read it looking for confirmation.
• Format and typing are read as proxies. They are not the point and everyone knows they are not the point, but a script with wrong margins, a paragraph of eight lines of action and three typos on page one signals a writer who has not read many screenplays, and the reader adjusts their expectations accordingly. This is unfair and it is also true, and it costs you nothing to remove.
• Action lines are read at speed and sometimes skipped. Write them so they can be skimmed without losing the story: short paragraphs, one image each, the important thing first in the sentence.

There is an exercise that is worth more than any amount of advice about this. Write the coverage of your own script as a hostile reader would write it: the logline they would write, the synopsis, the four grid ratings, and a comments section that names your script's weakest element in one sentence. If you cannot bring yourself to write the negative sentence, you already know what it says.

The logline, as an instrument rather than a summary

Lesson 2 used the logline as a diagnostic: if you cannot write it, the premise is unbuilt. Here it is a different tool. This one has to make a stranger want to read a hundred and ten pages, in one sentence, usually under forty words.

A working logline carries five things and drops everything else: a protagonist described by the trait that makes this story hard for them, an inciting event, a goal, the force opposing it, and something ironic or specific enough that it could only be this film. It gives away the setup and withholds the ending.

Draft	What is wrong
A woman returns to her childhood home after her father dies and confronts the past.	No goal, no opposition, no clock, no irony. True of four hundred films.
Kate Mahoney, estranged from her father since the incident at the co-operative in 2003, inherits Ashfield Farm and must deal with Derek Aylward.	Names nobody knows, an incident nobody can picture, and a verb, deal with, that means nothing.
A stunning, emotionally devastating story of loss and redemption set against the beautiful backdrop of rural beekeeping.	Adjectives doing the work of events. The writer is telling the reader how to feel instead of showing them what happens.
A woman with three weeks to sell her dead father's bee farm before the bank takes it must learn to keep the hives alive from the neighbour her father ruined.	Works. Protagonist with a defining pressure, a clock, a concrete goal, an obstacle who is a person, and an irony: the only help available is the person with most reason to refuse.

Note what the good version does not contain. No names, because a stranger cannot attach to a name they have never heard. No theme, because the theme is what the film discovers. No ending, because the pleasure of asking whether she can do it is the entire product you are selling.

Remember: The logline is not a compressed synopsis. It is the question your film asks, phrased so that a stranger wants the answer.

The pitch, out loud, in a room

A pitch is a performance with a time limit and an audience of people who have already read your logline and are deciding whether to spend money or time. Five minutes is a lot. Two is often enough.

1. The hook, one sentence. Essentially the logline, spoken as though you thought of it that morning.
2. Why this, why now, two sentences. What the film is about underneath, or what makes it timely, or what makes it cheap, or what makes it a part an actor would want. Choose the one true thing, not all four.
3. The story in three moves. The situation and the disruption; the middle stated as escalation rather than as a list of scenes; the crisis. Keep it in present tense, and keep the number of proper nouns as low as you can bear.
4. The ending. Tell them. Withholding the ending in a pitch is the single most common beginner error. You are not selling a mystery to a customer; you are demonstrating to a collaborator that you know how it finishes.
5. Why you, one sentence. Not a biography. The reason this particular story is yours to write.

Two practical matters. First, you will be interrupted, usually in the middle of move three, and the interruption is good news, because a person who is bored does not interrupt. When it happens, answer the question fully and then say where you were: after that she photographs the register, and here is what it costs her. Do not restart, and do not press on as though nothing was said.

Second, the comparison pitch, of the form this film meets that film, is efficient and dangerous. Efficient because Lesson 10 showed how much a phrase like Jaws in space transmits at no cost. Dangerous in three ways: comparing yourself to a masterpiece invites the obvious verdict, comparing yourself to a flop attaches its numbers to you, and a comparison that does not fit sends the listener off in a direction you then have to argue them back from. If you use one, make both halves films the room has certainly seen, and make the join do real work.

The written approach

Most scripts arrive by email rather than in a room, which makes the query a short, dull, disciplined document. One paragraph naming the title, form, genre and logline. One paragraph on you, containing only facts: a competition placing, a produced short, a relevant job, or nothing at all, which is fine. One line asking whether they would like to read it. No adjectives about your own work, because a writer who calls their script gripping has told the reader that they have not been in this position before, and no explanation of the theme.

On protection: in the United States copyright exists automatically when the work is fixed, and registration with the Copyright Office adds legal remedies. The Writers Guild also runs a registration service, which is evidence of a date rather than a copyright. Neither prevents anybody from writing a film about a beekeeper. What is protected is your expression, not your idea, and the practical risk to an unknown writer is being ignored rather than being stolen from.

Where scripts get read without an agent

The honest answer is that the routes are few and competitive, and they are also real.

• Competitions. The Academy of Motion Picture Arts and Sciences runs the Nicholl Fellowships in Screenwriting, an open competition restricted to writers who have not earned above a stated sum from screenwriting, and placing in it is read as a genuine signal.
• The Black List survey. First published in 2005 by Franklin Leonard, it is an annual list compiled from the votes of development executives naming the most liked screenplays not yet produced. It is a barometer of what is circulating and of what a good script that has not been made looks like.
• Labs and schemes. Sundance Institute runs screenwriting labs, and the British Film Institute publishes and supports new work through its funding and its programmes. These select for voice and development potential rather than for a finished commercial object.
• Shorts and festivals. Lesson 12 covered why. A finished eight minute film is a stronger calling card than a hundred and ten unread pages.

Worth holding on to: Everything in this lesson is downstream of the script being good. These are the mechanisms by which a good script gets found, and they cannot manufacture the thing they are looking for.

Common misconceptions

• A pass means the script was read carefully and rejected on the merits. A pass is the default outcome of a system with far more scripts than capacity. It is information about a system, and only sometimes about your script.
• You should withhold the ending in a pitch to create intrigue. The room is deciding whether you can finish a film, and the ending is the proof. Withholding it reads as not having one.
• Format is a trivial matter of taste. It is trivial in itself and it functions as a signal about how much you have read, which changes how generously page sixty gets read.
• A logline should summarise the plot. A summary tells the whole story badly. A logline poses the question the film answers and stops.
• Ideas get stolen. Copyright protects expression, not premises. For an unknown writer the realistic risk is silence, not theft.

The short version

• Coverage is a two page document with a logline, a synopsis, comments, a grid and a verdict of pass, consider or recommend, and the executive usually reads it instead of your script.
• A verdict forms early and the rest of the read confirms it, so the first ten pages set the frame and clean format removes a free reason to doubt you.
• Write your own hostile coverage. The sentence you cannot bring yourself to type is the note.
• A logline carries a pressured protagonist, an inciting event, a goal, an opposing force and an irony, in under forty words, and it never gives the ending.
• A pitch runs hook, why this and why now, three moves, the ending, and one line on why you. Expect interruption and treat it as interest.
• Comparison pitches transmit a great deal cheaply and attach whatever else the compared films carry.
• Copyright protects expression rather than ideas, and the realistic risk to a new writer is being ignored.

Sources

1. Writers Guild of America West. (n.d.). Writers Guild of America West. WGAW. wga.org
2. Academy of Motion Picture Arts and Sciences. (n.d.). The Academy. Oscars. oscars.org
3. United States Copyright Office. (n.d.). U.S. Copyright Office. copyright.gov
4. British Film Institute. (n.d.). Sight and Sound. BFI. bfi.org.uk
5. Goldman, William. Which Lie Did I Tell? More Adventures in the Screen Trade. Pantheon, 2000.

What five months on a picket line was about

From May to September 2023 the Writers Guild of America was on strike, and American film and television production largely stopped. Look at what the dispute was about, because the list is a description of the job itself. Residuals from streaming, meaning payment for reuse of work already delivered. Minimum staffing and duration for television writing rooms, meaning whether the work exists as employment at all. And terms governing generative artificial intelligence, meaning who counts as the author of a script.

Nobody strikes for five months over a hobby. Screenwriting is a trade with a labour history, a bargaining agreement, an arbitration system for credit and a long argument about reuse, and the reason to know that at the start of your career is that it tells you what shape the work has. It is not a salary. It is a sequence of engagements, paid in lumps, with long unpaid intervals in which the next thing is made.

In short: The economics are episodic. Everything in this lesson about daily practice follows from the fact that nobody is going to notice whether you wrote today.

Where the money actually comes from

Route	What it is	What it is like in practice
Spec sale	You write a script on your own time and somebody buys it outright.	The romantic route and the rarest. Most specs sell nothing and function as samples that get you considered for other work.
Option	A producer pays for the exclusive right, for a fixed term, to try to set the film up.	Usually a small sum against an agreed purchase price. Options lapse constantly and can be renewed. Being optioned is not being made.
Assignment	You are hired to write something somebody else controls: an idea, a book, a remake, an existing property.	Where most professional income comes from, and you get it because somebody read a sample they liked.
Rewrite and polish	You are hired to work on an existing script, often one you did not originate.	Common, well paid relative to the time, and frequently uncredited under Guild credit rules.
Residuals	Payments for the reuse of work already delivered, negotiated collectively.	The reason a credit from years ago can still pay something, and the reason reuse terms are worth striking over.
Prizes, labs and schemes	Competitions, fellowships and development programmes.	Rarely enough money to live on. Valuable because they cause your script to be read by people who otherwise would not.

Commissioned work is usually structured as a step deal: payment at each stage, treatment, first draft, revision, polish, with the employer holding the right to stop between steps. That structure explains something that upsets new writers: you will be replaced on projects, and you will be hired to replace others. It is not a verdict on your talent. It is what a step deal is for.

William Goldman, who wrote Butch Cassidy and the Sundance Kid and All the President's Men, gave the industry its most quoted sentence in Adventures in the Screen Trade in 1983: "Nobody knows anything." He meant that nobody can reliably predict what will work, which is a genuinely useful thing to hold onto, because it means a rejection is a prediction and predictions in this business are poor.

A practice measured in pages

Here is the part of this course that decides whether you ever finish anything. Screenwriting is unusual among the arts in that the barrier is almost never ability. It is the fact that nothing outside you requires the pages to exist.

1. Set a quota in pages, not hours. Hours reward sitting. Two pages a day finishes a feature draft in eight weeks including bad days. Pick a number you can hit on your worst day, not your best, because the purpose of the quota is that it survives a bad week.
2. Never stop at the end of a scene. Stop mid-scene, mid-exchange, ideally mid-line, when you know exactly what comes next. Starting tomorrow then costs nothing, whereas starting at a blank scene heading costs an hour and sometimes a week.
3. Keep the notebook separate from the draft. When an idea arrives for page ninety while you are on page forty, write it in the notebook and keep going. Writers who chase forward ideas produce four brilliant openings a year and no drafts.
4. Do not rewrite while drafting. Lesson 13 is the reason: passes come after a complete draft exists, and revising forward means polishing scenes that structure will later delete. Write badly and knowingly, and fix it in pass 2.
5. Put a deadline outside your own head. A person expecting pages on Sunday is worth more than any amount of resolve. A writing group, a reading partner, a competition deadline, a director who wants it: any external date will do, and none of your internal ones will.
6. Track days worked, not quality. A mark on a calendar for each day you hit the quota. You cannot judge quality while writing and you can count days perfectly, so count the thing you can count.

What about the days when nothing comes? Most of what gets called writer's block is not a mysterious affliction; it is a structural problem wearing a costume. You do not know what happens next because the outline is wrong, or because the scene you are trying to write has no want in it, or because you already know on some level that the last ten pages went the wrong way. Treat the blank afternoon as a diagnostic signal and go back to the step outline from Lesson 7 rather than sitting there harder.

Bottom line: The writers who make careers are not the most talented people in their year group. They are the ones who finished, then finished again.

The arithmetic of a career

Work out the timescale honestly. Outlining takes weeks. A first draft at two pages a day takes about two months. Passes take another two to three. That is roughly six months to a script you would show people, which means one, or with discipline two, features a year, and it means that a decision to write something is a decision about half a year of your life.

Which is why working writers keep three slots running. One script being written. One finished and out with people, which is generating waiting rather than work. And one being thought about, read around and outlined in a notebook. The three-slot habit is the practical answer to the emotional problem of submission: when a pass arrives on the script in slot two, you are already in the middle of the one in slot one, and the pass is an event in a week rather than a verdict on a life.

Rejection, counted

You control two numbers: how many scripts you finish and how many people read them. You do not control whether any given reader likes one. So keep a submission log with the date, the recipient, the script and the outcome, and look at it monthly. It converts a formless dread into a small table, and small tables have obvious next actions: this month nothing went out, so send something.

The second script matters more than the first. A first script that works can be luck, and everyone in the industry knows it; a second one that works is a writer. So the correct response to a good response on your first script is not to wait for the phone. It is to be already thirty pages into the next thing when the phone does or does not ring.

Choosing what to write next

Two filters, applied together. First, what can you not stop thinking about? Six months is too long to spend on a story you are merely willing to write, and readers can feel the difference between conviction and competence in about ten pages. Second, can it be made at a scale somebody might plausibly fund near you? A film with four characters, one strong location and a real want in it is not a lesser ambition than a film with a war in it. It is the one that gets shot.

And write the sample that shows what you would be hired for. If you want to write intelligent thrillers, the script that gets you hired to write intelligent thrillers is an intelligent thriller, not a proof that you can do anything.

So what?: Pick the thing you cannot stop thinking about that could actually be shot, and give it half a year on purpose rather than by accident.

Common misconceptions

• Writers wait for inspiration. Working writers keep a quota and treat inspiration as a bonus that occasionally turns up during the session it would never have turned up outside.
• Writer's block is a condition. It is usually a diagnosis: a wrong outline, a scene with no want, or ten pages that went the wrong way. It responds to structural work, not to willpower.
• Being rewritten means you failed. The step deal is built for replacement, credit is arbitrated precisely because multiple hands are normal, and rewriting others is itself a large part of professional income.
• The first sale is the finish line. An option is not a production, a sale is not a film, and most optioned scripts are never made. Plan for a body of work rather than an event.
• You have to be somewhere specific. Scripts travel, competitions and labs read from anywhere, and film cultures with public funding operate on different routes entirely. Where you live changes the route rather than the requirement, which is finished pages.

Looking back

• Screenwriting income is episodic: specs, options, assignments, rewrites and residuals, usually structured as step deals with a right to stop.
• The 2023 strike was about reuse payments, whether the work exists as employment, and who counts as an author. All three are permanent features of the trade.
• Set a page quota you can hit on a bad day, stop mid-scene, keep a separate notebook, and do not rewrite while drafting.
• Put the deadline outside your own head and count days worked rather than judging quality while working.
• Blocked usually means the outline is wrong, so go back to the step outline instead of sitting there harder.
• Run three slots: one being written, one out with people, one being thought about.
• Count the two numbers you control, finished scripts and readers reached, and start the next script before the answer arrives on the last one.
• Choose what you cannot stop thinking about, at a scale that could be shot, in the form you want to be hired for.

Sources

1. Writers Guild of America West. (n.d.). Writers Guild of America West. WGAW. wga.org
2. Writers Guild of America East. (n.d.). Writers Guild of America East. WGAE. wgaeast.org
3. Wikipedia contributors. (2024). 2023 Writers Guild of America strike. Wikipedia. en.wikipedia.org
4. British Film Institute. (n.d.). Features and interviews. BFI. bfi.org.uk
5. Goldman, William. Adventures in the Screen Trade. Warner Books, 1983.