A scene isn't a location. It isn't a conversation. It's a small machine built from three moving parts — and if any part is missing, the machine doesn't run.
Phase 1 · Screenwriting Foundations · Week 13 of 32For eight weeks you've been building at the macro level — premise, theme, character, world, structure. Everything has been documents about scenes: outlines describing what happens, promise maps listing moments the audience should experience, escalation ladders charting the antagonist's campaign. None of it has been a scene. This week, you write scenes. Not polished ones. Not connected ones. Not scenes that will necessarily survive into the final draft unchanged. Five standalone scenes drawn from different parts of your outline, written rough and fast, for one purpose: to learn the engine. The scene engine is the micro-structure that operates inside every working scene in every feature screenplay ever written: a character enters with a goal, encounters friction that prevents easy achievement of that goal, and the scene ends on a turn that changes the situation in a way that propels the story forward. Goal. Friction. Turn. Three parts. If any one is missing, the scene is dead weight.
Goal. Every scene begins with a character who wants something. Not a grand, philosophical want — not the protagonist's overarching Want from the character dossier — but a specific, immediate, scene-level goal. "Get the test results." "Convince the principal to delay the board vote." "Find out if the colleague knows." "Get through dinner without revealing what she discovered today." The scene-level goal may align with the character's overarching Want or it may conflict with it — sometimes the protagonist's scene goal is to avoid dealing with the very thing their Want demands. Either way, the goal must be present. A character who enters a scene without wanting anything specific is a passenger. The scene will move around them rather than through them.
The goal doesn't need to be stated. Often it shouldn't be. The audience infers the goal from context and behavior. If Nora walks into the principal's office after discovering the contamination data, the audience knows she's there to report — or to decide whether to report. Her goal is legible from the situation. But the writer must know the goal, even if the character doesn't announce it, because the goal determines what happens in the scene: what the character says, what they hold back, where they direct their attention, and how they respond to obstacles.
Friction. Friction is whatever prevents the character from achieving their scene-level goal immediately and easily. If the character enters wanting something and gets it in two lines of dialogue, there's no scene — there's a transaction. Friction creates the scene's dramatic content: the negotiation, the confrontation, the improvisation, the suppression. Friction comes in five varieties, and the strongest scenes use more than one simultaneously.
External opposition: another character actively resists. The principal doesn't want to hear the report. The colleague lies. Graham refuses to open the files. This is the most visible form of friction — two people in a room with incompatible goals.
Informational friction: the character doesn't have what they need — the right question, the right evidence, the right leverage. They're fishing, probing, trying to extract information from someone who holds it. The friction is the gap between what they know and what they need to know.
Social friction: the character can't pursue their goal openly because the social context forbids it. They want to confront someone but they're at a dinner party. They want to reveal evidence but the wrong people are in the room. The friction is the gap between what the character wants to do and what the situation allows them to do.
Internal friction: the character's own Wound, fear, or doubt resists the goal. They came to report the contamination but can't bring themselves to say the words, because the last time they reported something, the consequences destroyed colleagues they cared about. The friction is the gap between what they intend and what their psychology allows.
Temporal friction: a deadline, an interruption, a ticking clock. The character has five minutes before the meeting ends. The phone is about to ring. The other person is already standing to leave. Time is the obstacle, and the character must compress their goal into a narrowing window.
Turn. The turn is how the scene ends — the event, revelation, decision, or shift that changes the situation. A turn isn't a conclusion (the conversation ends normally, everyone goes home). A turn is a change. Something is different after the scene than it was before. The protagonist knows something new. The power dynamic has shifted. A door has opened or closed. A relationship has been damaged or forged. An option has been gained or lost. The turn is the scene's contribution to the larger story: the reason this scene exists in the screenplay rather than being summarized in a transition or skipped entirely.
The turn should land at or very near the scene's end. As you learned in Week 2: end one beat after the thing that changes. Don't let the scene continue past the turn — don't write the reaction shot, the summarizing dialogue, the emotional processing. Let the audience carry the implication of the turn into the next scene. The cut after the turn is where the audience does their work.
The engine in action. When all three parts are working, the scene has a felt shape: entry with energy (the goal creates forward motion), escalation through resistance (the friction creates texture and tension), and exit with consequence (the turn creates propulsion into the next scene). The audience experiences this as "a scene that works" — but what they're actually experiencing is a micro-story: beginning (goal), middle (friction), end (turn). Every scene in your screenplay should be a story this compact and this complete.
Common engine failures. When a scene feels flat, the diagnosis is almost always one of three problems. Missing goal: the character enters the scene without wanting anything specific. The scene drifts — dialogue happens, information is exchanged, but there's no forward motion. Fix: give the character a concrete objective before writing the scene. Insufficient friction: the character gets what they want too easily. The scene is a transaction, not a negotiation. Fix: add at least one form of friction — ideally two types operating simultaneously. Missing turn: the scene ends without anything changing. The situation after the scene is identical to the situation before it. Fix: ask "what's different after this scene?" If the answer is "nothing," the scene hasn't earned its place.
This week you're writing your first actual scenes with dialogue. The formatting of dialogue in a screenplay is standardized and non-negotiable — not because formatting is precious, but because every reader in the industry processes dialogue according to the same visual conventions. Deviation signals amateurism and slows the read.
The conventions: the character name appears centered (or indented to approximately 3.5 inches from the left margin), in ALL CAPS. Dialogue appears below the name, indented to approximately 2.5 inches from the left margin with a right margin of approximately 2 inches — narrower than action lines, creating the distinctive "dialogue column" that makes screenplay pages visually distinctive. Parentheticals, when used (sparingly — see Week 3), appear on their own line between the character name and the dialogue, indented slightly further than the dialogue.
A common formatting question: when a character's dialogue is interrupted by action and then continues, use (CONT'D) after the character name on the resumed speech. If a character's dialogue continues on the next page, use (MORE) at the bottom of the page and (CONT'D) at the top of the next — though most screenwriting software handles this automatically.
One practical note for this week: don't worry about perfect formatting in your rough scenes. Industry-standard screenplay formatting software (Highland, WriterSolo, Fade In, or the free web tool Fountain) handles margins, indentation, and wrapping automatically. If you're writing in a plain text editor, approximate the layout. The point this week is the engine, not the formatting. You'll have time to polish formatting before the draft.
10-minute drill: Write a single-page scene — three to four exchanges of dialogue between two characters, with at least three action lines interspersed. Format the dialogue manually (if not using screenwriting software): character names centered and capped, dialogue indented and narrower than action, parentheticals used only if absolutely necessary (aim for zero). Focus on making the dialogue column visually distinct from the action lines. Total output: one formatted page. Time yourself.
Assignment: Read 2 produced feature screenplays, chosen for scene-level construction.
Script Pair Brief: Select one screenplay known for lean, propulsive scenes — a script where every scene enters late, exits early, and wastes nothing — and one known for longer, more expansive scenes that build tension through accumulation and sustained duration. The first teaches you the engine at maximum compression: how a scene can establish a goal, create friction, and land a turn in a page and a half. The second teaches you what happens when the engine runs at slower RPM: how a five-page scene sustains tension through layered friction, subtext, and the controlled delay of the turn.
Where to find scripts legally: IMSDb, The Script Lab, SimplyScripts, studio FYC releases, and public library systems.
Why this pairing: Scene pace is a stylistic choice with structural consequences. Lean scenes create rapid momentum — the reader whips through pages. Extended scenes create depth — the reader sits inside a situation and feels its pressure accumulate. Your screenplay will need both. The pairing teaches you the range of the engine: how compressed it can be and how much it can sustain without stalling.
Reading Lens (track these while reading):
1. For at least 5 scenes in each screenplay, identify the three engine parts: what goal drives the scene, what friction resists the goal, and what turn changes the situation? 2. How late does the scene enter — how close to the first beat of friction does the writer begin? How early does it exit — how quickly after the turn does the writer cut? 3. How many types of friction operate simultaneously in the strongest scenes? Find a scene with at least two types working at once and note how they interact. 4. Find a scene where the turn is not what the character intended — the character entered with one goal and the scene ended by changing something completely different. How does that redirection work? 5. Identify the weakest scene in each screenplay — the one where the engine is least visible. What's missing: goal, friction, or turn?
Journal Prompts:
1. Choose the best scene from each screenplay — the one where the engine runs most powerfully. Write a one-paragraph analysis naming the goal, the friction type(s), and the turn. What makes this scene work better than the surrounding scenes? 2. For the lean-scene screenplay: find a scene that accomplishes its goal in under two pages. How much setup does the writer skip? What does the audience infer without being told? 3. For the extended-scene screenplay: find a scene that runs four or more pages. Map the friction escalation within the scene — where does the tension increase, plateau, and spike? How does the writer prevent the long scene from stalling? 4. Find a scene where the friction is primarily internal — the character is fighting themselves rather than another person. How does the writer externalize the internal conflict so the reader can see it? 5. Compare the scene-ending techniques: how does each screenplay handle the moment after the turn? Does the writer cut immediately, or allow a beat of reaction? Which approach is more effective in each case?
Deliverable: 5 standalone scenes from your feature screenplay.
Constraints: Write five scenes from different parts of your 8-sequence outline. They don't need to be consecutive. Choose scenes that represent a range: at least one from the first half and one from the second half, at least one confrontation scene (protagonist vs. antagonist or antagonist's agent), at least one scene driven primarily by internal friction, and at least one scene that delivers a promise-map item. Each scene should be 2–4 pages long, fully formatted with scene headings, action lines, and dialogue. Before writing each scene, note the engine on a separate line at the top: GOAL: [what the character wants in this scene], FRICTION: [what prevents it, and what type], TURN: [what changes by the end].
These are rough scenes. Do not polish. Do not rewrite dialogue for perfection. Do not worry about whether the scene will survive into the final draft. The purpose is to practice the engine — to feel what it's like to write a character pursuing a goal through friction toward a turn. You'll rewrite two of these scenes next week with a focus on subtext. For now, write them raw.
Quality bar: Every scene must have all three engine parts visibly working. If you write a scene and realize the goal is missing, add one before moving on. If you write a scene and the character gets what they want without resistance, add friction. If you write a scene that ends without anything changing, find the turn. The scenes should demonstrate that you can hear your characters — that the dialogue sounds like people in rooms, not like plot-delivery mechanisms. If every line of dialogue advances the plot and no line reveals character, the dialogue is doing one job when it should be doing two.
Estimated time: 6–8 hours (1–1.5 hours per scene, plus reading time).
Human Draft Reminder: You write the pages. AI helps you think, test, and decide.
For the first time, the Two Readers are evaluating screenplay pages — actual scenes with dialogue and action, not concept documents. This is a different kind of feedback. Reader A still evaluates structural function (does the engine work?). Reader B evaluates the reading experience (does the scene pull me in?). Write all five scenes before running the prompts. Choose the two scenes you're most uncertain about — the ones where the engine might not be working — and submit those. Save the others for next week's subtext revision.
Add this week's entry to your Disagreement Log. Scene-level feedback exposes a new kind of divergence: Reader A may find the engine mechanically sound (goal clear, friction active, turn present) while Reader B finds the scene emotionally inert or the dialogue indistinguishable between characters. Conversely, Reader B may be charmed by a scene's voice and texture while Reader A notes that the scene has no turn — nothing changes, and the scene could be cut without affecting the story. Both notes target real problems. The fix for a structurally sound but emotionally flat scene is usually specificity — more particular dialogue, more physical detail, more friction types running simultaneously. The fix for an emotionally alive but structurally loose scene is usually the turn — find the change that makes the scene consequential. Log the disagreement. Note which scene needs which fix.
Budget tier this week: Your chosen tier, now tested at the scene level.
Top 3 cost drivers in scene writing: 1. Scene count — every scene heading is potentially a new camera setup. A screenplay with 80 scenes costs more to shoot than one with 55 scenes, because each setup requires lighting, blocking, and rehearsal time. When two scenes in the same location could be combined into one longer scene with the same dramatic content, the combined version is cheaper and often more dramatically effective. 2. Characters per scene — every character in a scene is an actor on set. A scene with five speaking characters requires five schedules to align, five performances to capture, and five times the coverage (camera angles). Can the scene achieve its goal with fewer characters? A two-person confrontation is almost always more dramatically focused and cheaper to shoot than a five-person committee scene that delivers the same information. 3. Props and set interaction — a scene where the character opens a file, reads a document, checks a phone, and examines lab equipment requires each of those props to be sourced, placed, and continuity-tracked across takes. A scene where the same information arrives through dialogue requires none of that. The trade-off: dialogue-only scenes are cheaper but visually static. Action-with-props scenes are more cinematic but more expensive. Choose based on what the scene's friction type demands.
Cheaper equivalent: A three-scene sequence where the protagonist visits three different locations to gather three pieces of information can be compressed into one scene where all three pieces arrive in the same room — through a phone call, a document, and a conversation, stacked in rapid succession. Three setups become one. The scene is denser, cheaper, and arguably more tense because the information accumulates without relief.
Worth-it spend: Your confrontation scene — the one where protagonist and antagonist share the screen. This scene needs to breathe. It needs physical space, specific props, and enough shooting time for the actors to find the subtext. A confrontation scene shot in a rush, in a cramped space, with no room for behavior between lines, will feel like a TV scene dropped into a feature. Give it room. Give it objects. Give it silence.
A professional script reader forms their judgment of a writer's skill at the scene level — not the structural level. Structure can be fixed. An outline can be reorganized. But the ability to write a scene that breathes — where the dialogue sounds human, the action lines are visual and specific, and the turn lands with felt consequence — is the skill that separates writers who get meetings from writers who get form rejections. Every scene in your screenplay is an audition. The reader is asking: does this writer know how to make a scene work? Your five rough scenes this week are your first opportunity to prove that you do.
Writing scenes is a fundamentally different experience from writing documents about scenes. For the first time in this curriculum, you had to make your characters talk — and the gap between how you imagined them speaking and how they actually sounded on the page may have been disorienting. Write about that gap. Which character surprised you by sounding different than you expected? Which one resisted you — the character whose voice you couldn't find, whose dialogue felt generic or forced? That resistance is diagnostic. It may mean the character's design needs more specificity. Or it may mean the character's voice exists and you haven't found it yet — and the search is part of the drafting process.
By the end of this week you should have:
• Read 2 produced screenplays with a focus on scene-level construction (lean vs. extended)
• Written 5 standalone scenes from your feature (2–4 pages each, rough, with engine notes)
• Demonstrated all three engine parts (goal, friction, turn) in every scene
• Used at least two types of friction simultaneously in at least one scene
• Completed the dialogue formatting drill (one page of formatted dialogue, 10 minutes)
• Submitted 2 scenes to both Reader A and Reader B prompts
• Updated your Disagreement Log with this week's entry
Week 14 is Scene Engine II: Subtext, Status, and Silence — and it's the Phase 1 gate week. You wrote five scenes this week with the engine running on the surface: goals stated or clearly implied, friction visible, turns landing overtly. Next week, you learn to push the engine below the surface. What characters don't say. The power dynamics encoded in who sits, who stands, who speaks first, who breaks eye contact. The moments where silence does more work than any line of dialogue. You'll choose two of your five scenes and rewrite them with subtext operating beneath every exchange — so that the scene has two layers: the conversation the characters are having and the conversation the characters are actually having. The Phase 1 gate follows: concept doc, character dossier, antagonist plan, 8-sequence outline, and five scenes must all exist as documents before you proceed to drafting.