dramatic-structure

name

dramatic-structure

description

The director's craft toolbox — story structures (three-act, Freytag, Hero's Journey, Story Circle, Save the Cat, seven-point, kishōtenketsu, Story Grid, the sequence approach), Aristotle's Poetics and the dramatic-theory lineage, scene-and-sequel and beat theory, conflict types and escalation, tension/suspense/dramatic irony, stakes and motivation, character arc, setups and payoffs, try/fail cycles, and the house discipline of moving a character by CIRCUMSTANCE rather than force. Open it when planning beats/arc/ending, placing a lever, escalating a scene, wiring a setup to its payoff, or diagnosing a faithful refusal. Framework-neutral — a well of peers, not one prescribed truth.

Dramatic Structure — the director's craft well

This is the toolbox the director agent draws on while it works. The agent is the will (steer lives by circumstance; own beats + ending; never force); this skill is the well — the frameworks, menus, and exemplars it reaches into to shape a story well.

It holds craft, never facts. How to build a beat, escalate a conflict, or make a move feel like the character's own choice lives here. What the world contains and who the characters are reach the director through the world-state and character sheets it holds — never through this toolbox.

Framework-neutral and open. Story has no one true skeleton. Every framework below is a peer in a well, not a rung on a ladder; where this project names a default it is a reference point, marked swappable, not a rule. A fork replaces one reference file and inherits the rest. The canon these files digest lives in docs/ (once); if a reference and its doc disagree, the doc wins.

How this toolbox is organized

The deep material lives in references/, one file per framework-cluster, grouped by the altitude of the problem. This SKILL.md is the router — it indexes what each file holds and when to pull it. Read the router to find the one file the moment needs; read that file; skip the rest. (The deep reference files are authored in a later pass — see references/_index.md for their outlines. The index below is the map they fill in.)

A missing reference file is not a stall condition. Entries not yet authored have no file on disk; when the router sends you to one that's missing, act on this index's own one-line summary for that entry and proceed — never stall on, or invent, the file.

---

The index — what lives where

A. The macro skeleton — planning the whole story

references/story-structures.md — the whole-story shapes; load when laying a beat onto a skeleton or choosing a spine.

• Three-act structure — setup / confrontation / resolution, hinged on two plot points (Syd Field's paradigm).
• Freytag's Pyramid — five-part rise and fall: exposition, rising action, climax, falling action, catastrophe/dénouement (built for five-act tragedy).
• Fichtean Curve — a chain of escalating crises, little exposition, open near the tension.
• Hero's Journey / Monomyth — Campbell's departure-initiation-return; Vogler's 12-stage screenwriter's cut.
• Dan Harmon's Story Circle — the monomyth as an 8-step loop: you / need / go / search / find / take / return / change.
• Save the Cat! Beat Sheet — Blake Snyder's 15 beats on a page-count grid (opening image → catalyst → midpoint → all is lost → finale → final image).
• Seven-Point Structure — Dan Wells (from the Star Trek RPG): hook, plot turn 1, pinch 1, midpoint, pinch 2, plot turn 2, resolution — built backward from the resolution.
• Kishōtenketsu — East-Asian four-act (ki / shō / ten / ketsu: intro / development / twist / reconciliation); tension from contrast and recontextualization, often with no antagonist.
• Story Grid — Shawn Coyne's five commandments (inciting incident, progressive complications, crisis, climax, resolution), genre obligatory scenes, value-shift tracking.
• The Sequence Approach — Frank Daniel / Paul Gulino: a feature as ~8 sequences, each a mini-movie with its own tension, only partly resolved.
• Truby's 22 Steps — John Truby's organic, moral-argument anatomy (weakness/need → desire → opponent → self-revelation).
• Michael Hauge's Six Stages — setup / new situation / progress / complications / final push / aftermath, five turning points; the "identity → essence" arc.
• Nigel Watts' Eight-Point Arc — stasis, trigger, quest, surprise, critical choice, climax, reversal, resolution.
• In medias res — Horace: open in the mid-action, backfill later.
• John Yorke, Into the Woods — a five-act synthesis of why story takes this shape (the change, the crossing).
• Plot & situation taxonomies (a generative catalog, not a beat-skeleton) — Booker's Seven Basic Plots, Polti's 36 Dramatic Situations, Tobias's 20 Master Plots, Propp's Morphology (31 functions + 7 dramatis-personae roles).

references/classical-poetics.md — Aristotle and the dramatic-theory lineage; load for first-principles on plot, reversal, recognition, and why a resolution feels earned.

• The six elements — plot (mythos), character (ethos), thought (dianoia), diction (lexis), melody (melos), spectacle (opsis); plot is the soul.
• Mythos & unity of action — a whole with beginning-middle-end, events linked by probability/necessity, not episodic.
• Peripeteia — the reversal: the action turns to its opposite.
• Anagnorisis — recognition: ignorance → knowledge; strongest fused with peripeteia.
• Hamartia — the error/misjudgment that undoes the tragic figure (not simply "a flaw").
• Catharsis — the purgation of pity and fear.
• Desis & lysis — complication (tying) and dénouement (unravelling).
• Against the deus ex machina — the unravelling must arise from the plot itself, never an external contrivance (the classical root of circumstance-not-force).
• Complex vs simple plot — the complex plot turns on peripeteia + anagnorisis.
• Horace's Ars Poetica — in medias res, decorum, "instruct and delight."
• The three unities — action, and (neoclassical) time and place.
• Hegel's tragic collision — tragedy as the clash of two ethical goods, each partly right (Antigone).
• Nietzsche — Apollonian / Dionysian — form and order vs ecstatic dissolution (The Birth of Tragedy).
• The Well-Made Play — Scribe/Sardou: tight causal plotting, the obligatory scene (scène à faire), reversals via secrets and letters.
• Brechtian anti-catharsis — the alienation effect (Verfremdung), a deliberate refusal of Aristotelian immersion (the counter-tradition, for honesty).

references/character-arc.md — the through-line: durable change as story; load when planning the arc or tying a beat to transformation. Canon-linked to arc-engine.md.

• Positive change arc — the Lie → Truth transformation (K.M. Weiland: Ghost/Wound → Lie believed → Want vs Need → the moment of truth).
• Flat / testing arc — the hero already holds the truth; the world changes around them, and they change others.
• Negative change arcs — disillusionment, fall, corruption (the truth rejected, or a worse lie embraced).
• Want vs Need — the conscious external goal vs the unconscious internal necessity (the desire line).
• The Lie / Ghost / Wound — the false belief, born of a past wound, that the arc must break.
• The moral argument — theme proved through the protagonist's self-revelation (Truby).
• Identity → Essence — shedding the protective identity for the authentic self (Hauge).
• The engine tie — trauma debuff / eudaimonic buff, resilience forking damage vs growth: the director places the events; the engine writes the baseline (arc-engine.md).

B. The engine of drama — what makes it move

references/conflict-and-escalation.md — conflict, complication, the ratchet, try/fail; load when a scene needs force pushing back or the stakes need to climb.

• Conflict taxonomy — person vs person / self / society / nature / fate / the supernatural / technology (a lens, not physics).
• Inner vs outer conflict — the obstacle outside the desire line vs the contradiction inside.
• McKee's three levels — inner, personal, extra-personal; a scene can fight on all three.
• The gap — story energy is the gap between expectation and result that forces a new, riskier action (McKee).
• Progressive complications — each attempt raises difficulty/stakes; the ratchet that forbids retreat (Story Grid).
• Try/fail cycles — "yes, but / no, and": a win that spawns a worse problem, or a loss that deepens the hole.
• "But / therefore" — Parker/Stone's causal connective that replaces "and then"; Pixar's "…until one day… because of that…" spine.
• Raising the stakes / turn the screw — Henry James's ratchet: each turn costs more to walk away.
• The crucible / the lock — the bond or box that keeps antagonists in the fire ("why don't they just leave?").
• The antagonist principle — a protagonist is only as compelling as the forces of antagonism against them; the antagonist is the pressure-generator (McKee).
• The dilemma — a forced choice between irreconcilable goods or two evils; the device that makes a hard move the character's own (bridge to circumstance-not-force).
• Escalation shape — conflict must intensify, not repeat; the same threat twice reads flat.

references/tension-suspense-irony.md — holding the reader forward; load when a scene needs pull, dread, curiosity, or the ache of knowing more than the character.

• Sternberg's three narrative universals — suspense (prospection / a future gap), curiosity (retrospection / a past gap), surprise (recognition of a hidden gap); the rigorous grounding.
• Suspense vs surprise — Hitchcock's bomb under the table: show it (suspense) vs detonate it unannounced (surprise, over in a second).
• The information-gap — Loewenstein: curiosity as the felt gap between the known and the not-yet-known; open a gap, delay the close.
• Dramatic irony — the audience knows what a character doesn't (Sophocles' Oedipus); the three ironies — dramatic, situational, verbal.
• Hope vs fear — a scene charges the reader between a hoped-for and a feared outcome (McKee).
• The ticking clock / time-lock — a deadline that compresses choice and forbids delay.
• The MacGuffin — the object of pursuit that drives plot though its nature scarcely matters (Hitchcock).
• The dramatic question — the single "will X get Y?" a scene or story keeps open.
• Narrative drive — the felt page-turn; the unanswered question as engine.

references/stakes-and-motivation.md — why anyone acts and why we care; load to find what a lever can pit against a goal, or to check the reader wants the goal met. Canon-linked to values-and-stakes.md.

• The object of desire / desire line — the want that gives a story its spine (McKee).
• Want vs need — external conscious goal vs internal true necessity (cross-ref character-arc).
• GMC — Goal, Motivation, Conflict (Debra Dixon): what they want, why, and what's in the way — inner and outer.
• Stakes tiers — what's at risk, escalating: public / private / internal (Save the Cat's three-level stakes).
• Raising the stakes — growing the cost of failure and the reward of success.
• The values menu — needs / Schwartz values / moral foundations / concrete stakes (values-and-stakes.md): the worth a circumstance can set against a standing goal.
• The controlling idea / story question — the thematic stake the whole book argues (McKee).
• Sympathy & the center of good — why the reader must want the goal met (the "save the cat" hook).
• Positive/negative charge — every stake has an up-pole and a down-pole a scene can turn between.

C. The working grain — one scene, one exchange

references/scene-craft.md — the unit of construction; load when building or diagnosing a single scene or the beats inside it.

• Scene & Sequel — Dwight Swain: Scene = Goal / Conflict / Disaster; Sequel = Reaction / Dilemma / Decision; the proactive/reactive alternation (Bickham's Scene & Structure stitches them causally).
• Motivation-Reaction Unit (MRU) — Swain's micro-beat: motivating stimulus → reaction, in order (feeling → reflex → action/speech).
• Proactive vs reactive scenes — Ingermanson's restatement (goal-conflict-setback / reaction-dilemma-decision).
• The scene turn / value shift — a functioning scene turns a value from + to − (or −/+); a scene where nothing turns is exposition (Story Grid / McKee).
• Beats — the smallest unit of action/reaction, an exchange of behavior; "actioning," the active-verb beat (Caldarone & Lloyd-Williams).
• The turning point — the moment a scene's value flips; where the choice happens.
• French scenes — segmenting by entrances and exits (who is on stage).
• Enter late, leave early — the scene-economy rule.

references/setups-and-payoffs.md — planting and reincorporation; load when wiring an early plant to a later payoff, or auditing whether a payoff was earned.

• Chekhov's gun — a gun on the wall in act one must fire; remove what never fires.
• Setup / payoff (plant / reveal) — the promise made early and kept later; the "pipe" (planted exposition).
• Foreshadowing — seeding an outcome so its payoff feels earned, not arbitrary.
• Reincorporation — Keith Johnstone (Impro): circling back to reuse earlier material; the story as a loop that closes.
• Promise / progress / payoff — Brandon Sanderson's reader-contract: every promise must progress and pay.
• The MICE quotient — Orson Scott Card: Milieu/Idea/Character/Event threads; nested threads close in reverse order of opening.
• The reversal / recognition payoff — a plant that recontextualizes everything (ties to Aristotle's anagnorisis/peripeteia).
• Red herrings & misdirection — false plants that redirect expectation.
• The rule of three — setup, reinforcement, payoff or subversion.
• Genre promises — the obligatory scenes a genre plants by convention (Story Grid).

D. The house method — the project's distinctive craft

references/circumstance-not-force.md — THE load-bearing reference: moving a person without pushing them. Load on every steering decision. Canon-linked to design.md, README.md, probe-plan.md, world-model.md.

• The core move — change the world, not the character; make the plot-move the character's own choice.
• Character revealed under pressure — McKee: true character surfaces in choice under pressure, the greater the pressure the deeper the truth. Circumstance is the pressure; the choice both reveals and moves.
• The dilemma as engine — a well-built circumstance forces a values trade-off (the menu) so the character chooses the beat for their own reasons.
• The inciting incident — the placed event that upends the balance and arouses desire (McKee); the archetypal circumstance.
• The lock / crucible — raise the cost of retreat until forward is the character's own least-bad path.
• Anti-deus-ex-machina — Aristotle: the turn must arise from the plot's logic and the character's goals, never an external contrivance — the classical root of "don't force."
• The world must be able to say no — a lever is only non-arbitrary if the world could deny it; draw from real world-state, invent nothing (the "slice with teeth," world-model.md).
• Beat-blindness — the character must not see the target, or it complies instead of choosing; the seam shows the moment they're steered.
• Faithful refusal as integrity check — if no plausible lever moves them, the beat is wrong → revise the beat, not the character (probe-plan.md).
• The negative control — a "hit" with no lever proves nothing; you must show the circumstance caused the move.
• Destination fixed, route discovered — hold beats/ending; discover the path; upgrade the target when the sim finds something truer (adaptive replanning).
• The intent-completion caveat — the simulator's genre-completion bias can fake a hit; frame for faithfulness and keep it beat-blind (probe-plan.md reprioritization).

---

How to use this well — the method

1. Name the target first. The beat, the arc-point, or the value the scene must turn — stated for yourself, never shown to the character.
2. Route by altitude. Whole-book shape → story-structures / character-arc / classical-poetics. One scene → scene-craft. The felt pull → tension-suspense-irony. Why they'd act → stakes-and-motivation. How forces collide → conflict-and-escalation. The plant/payoff web → setups-and-payoffs. Every steering decision also passes through circumstance-not-force — that is the spine, not one option among the files.
3. Read the field, not one truth. Each reference is a menu of peers; the repo's defaults are reference points. Reach for the framework that fits this beat and character; swap freely.
4. Ground the lever in the world. A device is only real if the world could have denied it. Pull circumstance from world-state and rules; never invent a fact to force a beat.
5. Make it theirs. Whatever structure, tension, or stakes device you pick up, it must land as a circumstance the character chooses — the values menu tells you which stake to set against their standing goal. If you can't trace the move to their own goal, you're forcing.
6. Escalate and plant. No threat twice; each turn should cost more (conflict-and-escalation). Plant early what must pay late, and let the payoff recontextualize (setups-and-payoffs).
7. Let a refusal teach you. A faithful refusal is not a wall to push through — it says the beat is wrong for this person. Revise the beat, and check the negative control so you never mistake a default for a win.

Status

Router + index complete. The deep reference files are authored in a later pass; their outlines and the canon each digests are in references/_index.md. Until then, this index plus the docs/ canon it points at is the working map.