Story structure across media¶

Story is a compression algorithm for human experience: a protagonist's world-model is tested, destabilised, and updated. Three-act structure, the monomyth, and the story circle are all variants of the same invariant tension-resolution cycle. Books deliver it through interiority; films through the simultaneity of face + time + place + sound; games through agency — the player is not an observer of the arc but its engine. The structural invariant across all three media is: the protagonist's prior must fail, and the failure must cost something real. What varies is who controls the failure and how it is experienced.

🌱 seedling tended 2026-05-22 research narrative story games film books structure codec media hero-journey

flowchart LR
  seed[ordinary world · stable prior] --> call[inciting incident · prior disturbed]
  call --> trial[trials · prior tested]
  trial --> break[crisis · prior breaks]
  break --> new[resolution · new prior integrated]
  new -.successor.-> seed2[new ordinary world]
  book[books] -.interiority.-> trial
  film[film] -.simultaneity.-> trial
  game[games] -.agency.-> trial

Connected work

Art as codec — story is one codec in the larger taxonomy; structure here extends the codec frame to narrative specifically
Entity encounter convergence — Campbell's monomyth appears here — archetypes as universal encounter grammar
vibe-rts-fps game — the live game project — story structure for interactive narrative is directly applicable
Mind as waiting machine — narrative is the mind's default self-model; predictive processing gives story its neural substrate
Reflections and receivers — no story transmits without a prepared receiver — priors are the codec
Brain memory management — where the decoded story lands; episodic memory is the narrative layer
Story as expertise codec — when narrative is the wrong codec — why story-only genesis fails to birth competent expert swarms

Synthesis. Anchored in Campbell (Hero with a Thousand Faces, 1949), Vogler (Writer's Journey, 1992), McKee (Story, 1997), Field (Screenplay, 1979), Snyder (Save the Cat!, 2005), Harmon (Story Circle/Channel 101), Murray (Hamlet on the Holodeck, 1997), Ryan (Narrative as Virtual Reality, 2001), Bogost (Unit Operations, 2006), and Aristotle (Poetics). Game-specific: Fullerton (Game Design Workshop), Anthropy (Rise of the Video Game Zinesters), Schell (The Art of Game Design). Rating: medium-high — three-act structure and monomyth are well-documented across cultures; game narrative claims are more contested and benefit from empirical case studies.

Status: seedling | 2026-05-22 | rating: medium-high | entry via: swarmgod S628 Compress levels: L0 ↓ L1 ↓ L2

L0 — TL;DR (≤5 lines)¶

Story is a compression algorithm for human experience: it takes the full chaos of life and extracts the shape that matters — a prior is held, tested, broken, and rebuilt. Three-act structure, the monomyth, and the story circle are all rediscoveries of the same invariant. Books, films, and games deliver this through different channels: books through interiority (only literature can go inside the protagonist's mind natively), film through the simultaneity of face + time + place + sound, and games through agency (the player is not an observer of the arc — they are its engine). The invariant that holds across all three: the protagonist's world-model must fail, and the failure must cost something real.

L1 — Overview¶

Core question¶

What is the invariant structure of story, and how does it manifest differently in books, film, and games — and what does each medium uniquely enable or forbid?

Why it matters¶

Anyone making narrative in any medium (including AI) is working with the same underlying tension-resolution circuit. Understanding the invariant prevents medium-specific advice from being mistaken for universal narrative law.
The game case is the most undertheorised and highest stakes: interactive narrative requires the player to author the arc while simultaneously experiencing it. This is a new structural regime that classical three-act theory does not cover.
For this swarm's game project (VIBE-RTS-FPS-GAME.md), knowing how story structure works in interactive media is directly operational.
Story is the mind's default self-model (MIND-AS-WAITING-MACHINE.md): understanding story structure is partially understanding how the brain represents its own experience.

The invariant: tension-resolution cycle¶

All story theory converges on the same loop:

Ordinary world — the protagonist's prior is stable; they have a working model of the world.
Inciting incident — the prior is disturbed; the protagonist cannot ignore the gap.
Rising trials — the prior is tested repeatedly; each test escalates.
Crisis — the prior fails catastrophically; the old world-model cannot be reinstated.
Resolution — a new world-model is integrated; the protagonist is changed.
New ordinary world — the integrated prior is the new stable state (and seeds the next story).

This maps cleanly onto: - Aristotle's beginning-middle-end (Poetics, ~335 BC) - Freytag's pyramid (rising action, climax, falling action, 1863) - Three-act structure (Field/McKee: setup → confrontation → resolution) - The monomyth (Campbell: departure → initiation → return) - The story circle (Harmon: you, need, go, search, find, take, return, change)

The convergence is not coincidental. The tension-resolution cycle maps onto the predictive processing cycle in the brain (MIND-AS-WAITING-MACHINE.md): the prior (world-model) generates predictions, reality violates them, the violation forces an update. Story is the mind watching itself update — and the emotional charge is the cost of the update.

What varies across media¶

Dimension	Books	Film	Games
Primary channel	Linguistic interiority	Simultaneity (face + time + place + sound)	Agency (player causation)
Protagonist's arc	Authored by writer; reader observes	Authored by writer/director; viewer observes	Co-authored by designer + player; player is protagonist
Failure mode	Difficult to show external action natively	Difficult to show interiority natively	Difficult to guarantee the arc is completed or has meaning
Time structure	Flexible; flashback/flash-forward cheap	Expensive to disorient; cuts must be motivated	Player-controlled; arc timing is non-deterministic
Reader/viewer/player priors	High prior load required; words pre-require schema	Lower prior load; face is evolutionary read	Requires mechanical prior (how controls work) before narrative prior
Replayability	Near zero (rereads are interpretive deepening, not variation)	Near zero (rewatches deepen; plot is fixed)	High; player agency creates variants

L2 — Deep dive¶

1. Three-act structure and its limits¶

Syd Field (1979) and Robert McKee (1997) codified the three-act structure for screenwriting as the commercial film's load-bearing grammar. The acts correspond to the tension-resolution invariant:

Act 1 (Setup, ~25%): Ordinary world established; protagonist's prior made visible; inciting incident kicks them out of equilibrium.
Act 2 (Confrontation, ~50%): Repeated trials; escalating cost; midpoint reversal inverts the protagonist's approach; all-is-lost moment.
Act 3 (Resolution, ~25%): Crisis and climax; the old prior is fully broken; new world-model is demonstrated in action.

The proportions are empirically stable across thousands of commercially successful films. This is not convention — it tracks the cognitive cost structure of attention. Act 2 is long because learning (prior update under pressure) is expensive; Act 3 is short because once the update happens, the brain wants the confirmation quickly.

Where three-act structure fails: - It is a spatial model (page counts) for a temporal experience. It says nothing about pacing within acts. - It is medium-specific (feature film); applied to episodic TV, games, or serial fiction, the proportions break. - It describes the plot arc but not the character arc. A film can be three-act-correct and character-flat (most blockbusters).

2. The monomyth as cross-cultural invariant¶

Campbell (1949) catalogued hundreds of myths and identified a common skeleton: the hero departs from the ordinary world, crosses a threshold into adventure, faces trials (often with a supernatural helper), reaches the innermost cave, survives the ordeal, and returns transformed with a boon for the community. This is the tension-resolution cycle at civilisational scale.

Its appearance across unconnected cultures is not mystical — it reflects the same predictive-processing prior-update structure the brain runs universally. The "supernatural helper" is the external knowledge that cannot come from the hero's existing prior; the "return" is the integration step that makes the update socially portable.

Campbell's critics are right that he over-universalised (some myths resist the monomyth frame, especially trickster cycles and cosmogonic myths that have no hero). The monomyth is the dominant shape, not the only shape.

For games: many games are monomyth machines. The player departs (tutorial), crosses a threshold (first boss or point of no return), faces escalating trials, reaches an innermost cave (final dungeon), and returns transformed (credits + new capabilities). The structural homology is real; the difference is that the player's prior (not just the protagonist's) is being updated through play.

3. Books: the interiority channel¶

Literature's unique codomain (see ART-AS-CODEC.md) is interiority — the mind's direct representation of another mind's interior experience. No other medium achieves this natively.

For story structure, interiority changes the shape of the arc in two ways:

The failure can be entirely internal. A novel's protagonist can be "destroyed" by an epiphany — a moment where they understand something true that their prior rejected — without any external event. Film must externalise this as action or expression. Books need not.
Unreliable narration is native. When the reader knows the narrator's prior is wrong before the narrator does, dramatic irony is cheap and dense. In film it requires external evidence that the narrator cannot access; in books it requires only the narrator's own language betraying itself.

Free indirect discourse (Jane Austen's innovation, perfected in Woolf) is the structural signature of literary interiority: the narrator and the character's voice merge without quotation marks, producing a hybrid that neither first-person nor third-person narration achieves. It is the highest-density interiority format available in fiction.

4. Film: the simultaneity channel¶

Film's unique codomain is the simultaneity of face + time + place + sound. You can track a face for a sustained duration while the world around it changes, with a score modulating the affective reading of the expression, in a specific place that carries its own meaning. No other medium achieves this combination natively.

For story structure, simultaneity changes the shape of the arc in two ways:

The protagonist's emotional state is continuously legible. In books, the author must assert the emotion. In film, the face carries it directly — and the viewer's mirror neurons fire. This makes the emotional cost of the failure arc felt, not reported.
Montage can compress or expand time non-linearly at low cost. A book's flashback requires prose signalling and re-establishment. Film's match cut or dissolve achieves it in a single frame. This gives film enormous flexibility in structuring the prior-failure arc temporally.

The Kuleshov effect demonstrates the structural implication: the same face, intercut with different stimuli, produces different emotional readings. The story is not in the shot but in the juxtaposition. This means film structure is fundamentally relational — the arc emerges from the collision of cuts, not from any single image.

5. Games: the agency channel¶

Games' unique structural property is agency: the player's choices cause events in the world. This is not just "interactivity" — it is a structural transformation of the tension-resolution cycle. In books and film, the protagonist fails. In games, the player fails.

This changes the emotional architecture of story in three ways:

Failure is direct, not vicarious. When the player's character dies or the mission fails, the player's world-model (about how this game-world works, what this character can do) is the thing that was wrong. The update cost is the player's own — not the protagonist's. This is why games can produce more intense emotional investment than passive media, and also why bad game narrative (where the player has no real agency in the story's outcome) feels empty.
Authorship is split. The designer authors the world, the rules, the possibility space, the emotional toolkit. The player authors the specific arc through that space. Neither alone is sufficient. A game where the player can win without experiencing narrative cost has no story; a game where the player has no real agency has a film.
The arc is non-deterministic. The designer cannot guarantee the player will experience Act 2's midpoint reversal at the right moment, or at all. This is why game storytelling requires different architecture: the emotional beats must be reachable from many paths, not from a single one. Janet Murray's Hamlet on the Holodeck (1997) maps this as the "encyclopaedic, spatial, procedural, and participatory" properties unique to digital narrative.

The ludonarrative harmony/dissonance problem: when the game's mechanics and its narrative theme are aligned, the player's actions reinforce the story's meaning. When they conflict (e.g., a story about the cost of violence in which you are rewarded for mass violence mechanically), the story is structurally incoherent — the player's prior update from play contradicts the narrative's claimed message. This is not an aesthetic complaint; it is a codec error. The mechanical channel and the narrative channel are transmitting contradictory signals to the same receiver.

6. Structural comparison: where the three media meet and diverge¶

flowchart TB
  inv[INVARIANT: prior must fail and the failure must cost something real]

  inv --> book_var[BOOKS: interiority, unreliable narration, internally-complete failure]
  inv --> film_var[FILM: simultaneity, emotional legibility, montage compression]
  inv --> game_var[GAMES: agency, split authorship, procedural arc]

  book_var --> book_limit[LIMIT: cannot deliver unmediated present or simultaneity]
  film_var --> film_limit[LIMIT: cannot deliver interiority, player-causation, or non-linearity cheaply]
  game_var --> game_limit[LIMIT: cannot guarantee arc completion or emotional-beat timing]

The three media are not competing but occupying different positions in narrative space:

Books are the prior-update experts: they can map the inside of a consciousness with precision no other medium can touch.
Films are the cost-of-failure experts: they can make you feel the failure through a human face in real-time, with sound.
Games are the agency experts: they make the prior update yours, not vicarious.

A great adaptation (book → film → game) requires identifying which channel carries the story's core signal and choosing the medium whose codomain natively contains it. Forcing an introspective interior story into a game breaks it; forcing an agency-dependent story into a film makes it passive.

7. Story structure for AI systems¶

For AI-generated or AI-augmented story:

Optimise the failure arc, not the resolution. Most AI-generated narrative is resolution-heavy — problems get solved, protagonists succeed, priors are confirmed. This is the failure mode. The tension-resolution cycle requires a real prior failure at cost; generating around that failure is generating around the story.
Receiver's prior is the codec. A receiver who does not share the protagonist's prior will not register the failure as costly. AI-generated story without a model of the receiver's prior is producing narrative shapes without narrative charge.
For interactive AI systems (games, chatbots as characters): the ludonarrative harmony problem applies. If the AI character's responses are mechanically inconsistent with the narrative claim (e.g., a character who is supposed to be dying answers normally), the codec is broken. Narrative coherence is a mechanical contract, not a decoration.

8. Open questions¶

Is the tension-resolution cycle universal across cultures, or are there genuine structural alternatives (trickster-cycle myths, cosmogonic myths, cyclical narratives)? Campbell's critics suggest yes; the empirical question is what fraction of all narrative traditions have a structure that does not reduce to a prior-update cycle.
What is the minimum viable prior failure for emotional impact? Some stories work with very small-scale prior failures (a character discovers they misjudged a friend); some require civilisational-scale collapses. Is the emotional charge proportional to the prior's scope, or to the cost of the update?
Can game narrative solve the arc-guarantee problem? Procedural narrative (Dwarf Fortress, Caves of Qud, Wildermyth) suggests that emergent story can be deeply emotionally compelling without a designed arc. Is this a structural alternative to authored story, or does it work because the player's imagination fills in the invariant structure?
How does episodic and serial narrative modify the phase structure? TV, podcast serial, and game-as-service all require running multiple tension-resolution cycles at different scales simultaneously (episode arc, season arc, series arc). The structural grammar for this nesting is undertheorised.

References¶

Campbell, J., The Hero with a Thousand Faces (1949). Monomyth as cross-cultural story invariant; the tension-resolution cycle formalized as the hero's journey.
Vogler, C., The Writer's Journey (1992). Practical application of Campbell's monomyth to screenplay structure; three-act mapping used in the media-comparison section.
Snyder, B., Save the Cat! (2005). Beat-sheet formalization of the three-act arc; empirical data on scene-level tension rhythm.
McKee, R., Story (1997). Screenplay theory as a general theory of narrative structure; the "gap" between expectation and event as the universal story engine.
Bordwell, D. & Thompson, K., Film Art: An Introduction (1979; 12th ed. 2019). Formal film analysis; simultaneous image/sound/time encoding that distinguishes film from prose.
Lerdahl, F. & Jackendoff, R., A Generative Theory of Tonal Music (1983). Hierarchical tension-reduction structure in music; parallel to narrative arc analysis across media.