# 05 — Story and Narrative Grammar

## 1. First principle: story is perceived change

A topic is what the content is about. A story is the ordered perception of change.

```text
STORY
= State A
+ Subject
+ Desire, need, question, or disequilibrium
+ Opposing force or uncertainty
+ Attempts, discoveries, or escalations
+ Irreversible change
+ State B
+ Meaning
```

A narrative is the controlled release of story information to an audience.

```text
NARRATIVE
= story facts
+ point of view
+ selection
+ order
+ duration
+ emphasis
+ withholding
+ revelation
+ audiovisual treatment
```

Therefore, the production system must maintain two separate graphs:

- **Story graph:** what happened, what is true, and what caused what.
- **Narrative graph:** what the audience is told, in what order, and with what treatment.

The narrative may reorder the story, but it may not silently falsify it.

## 2. Beat as the smallest story unit

```text
BEAT
= prior audience state
+ new information/action/emotion
+ turn
+ resulting audience state
```

Every beat has:

| Field | Question |
|---|---|
| Function | What job does this beat perform? |
| Proposition | What becomes known or believed? |
| Change | What is different after it? |
| Subject | Who or what carries the change? |
| Force | What causes or resists the change? |
| Stakes | Why does the change matter? |
| Evidence | What supports the proposition? |
| Emotion | What should be felt, and by whom? |
| Tension | What remains unresolved? |
| Dependency | What must the audience already know? |
| Payoff link | What earlier setup does it answer? |
| Setup link | What later beat does it prepare? |
| Duration | How much time does it require? |
| Treatment | Dialogue, scene, narration, graphic, sound, or combination? |
| Exit state | What question or expectation carries forward? |

## 3. Beat-function vocabulary

### Entry functions

```text
COLD_OPEN
HOOK
PROVOCATION
IMAGE_PROMISE
QUESTION_PROMISE
OUTCOME_PREVIEW
CONTRADICTION
MYSTERY
```

### Orientation functions

```text
CONTEXT
TIME_PLACE
CHARACTER_INTRODUCTION
RELATIONSHIP_MAP
DEFINITION
RULE_OF_WORLD
BASELINE
SCOPE
QUALIFICATION
```

### Motion functions

```text
INCITING_EVENT
GOAL_FORMATION
DECISION
ATTEMPT
PROCESS
DISCOVERY
ESCALATION
COMPLICATION
OBSTACLE
FAILURE
ADAPTATION
PROGRESS
REGRESSION
```

### Proof and reasoning functions

```text
CLAIM
REASON
MECHANISM
EVIDENCE
EXAMPLE
DEMONSTRATION
COMPARISON
CONTRAST
COUNTERARGUMENT
REBUTTAL
LIMITATION
UNCERTAINTY
SYNTHESIS
```

### Emotional functions

```text
VULNERABILITY
REACTION
RELIEF
DREAD
AWE
HUMOR
GRIEF
ANGER
HOPE
INTIMACY
REFLECTION
```

### Turn and payoff functions

```text
REVEAL
REVERSAL
RECOGNITION
CALLBACK
PAYOFF
CLIMAX
CHOICE
CONSEQUENCE
RESOLUTION
NEW_NORMAL
TAKEAWAY
CALL_TO_ACTION
TAG
```

One beat may hold multiple functions, but it should have one dominant function.

## 4. Causal grammar

A coherent sequence usually contains one or more causal chains:

```text
BECAUSE   A causes B
THEREFORE B changes the next choice
BUT       a force blocks the expected path
DESPITE   action continues under resistance
UNTIL     repetition reaches a threshold
UNLESS    a condition threatens the outcome
SO THAT   an action is tied to intent
WHICH MEANS a fact becomes relevant to the audience
```

Weak assembly often connects beats with only `AND THEN`. Strong assembly asks:

```text
Can adjacent beats be connected by BECAUSE, THEREFORE, BUT, or WHICH MEANS?
```

Use `AND THEN` deliberately for chronology, montage, accumulation, or absurdity—not as the default substitute for causality.

## 5. Information-state model

For every important proposition, track what the audience currently knows:

```text
UNKNOWN
TEASED
PARTIALLY_KNOWN
BELIEVED
QUESTIONED
CONTRADICTED
SUPPORTED
PROVEN_WITHIN_SCOPE
REINTERPRETED
RESOLVED
```

A reveal is only a reveal when the narrative state changes.

```text
REVEAL_STRENGTH
= importance
× prior uncertainty
× evidence credibility
× consequence
× preparation quality
- predictability
- confusion
```

## 6. Tension and open loops

```text
TENSION
= desired answer/outcome
- current certainty/control
+ cost of delay/failure
```

Open-loop types:

```text
question
promise
prediction
danger
missing cause
missing identity
missing method
unfinished action
moral dilemma
contradiction
countdown
comparison gap
```

Each open loop must have:

```text
setup_beat
importance
expected_resolution_window
status
payoff_beat or intentional_nonresolution
```

Rules:

```text
IF a loop is high importance AND remains unresolved past its expected window
THEN create a reminder, partial payoff, or explicit deferral.

IF more than the configured open-loop capacity is active
THEN close, merge, or deprioritize the weakest loops.

IF the payoff is smaller than the promise
THEN reduce the promise or strengthen the payoff.
```

## 7. Stakes grammar

Stakes answer, “What changes if this succeeds, fails, or remains unknown?”

Dimensions:

```text
holder: self | relationship | group | institution | public | environment
kind: physical | emotional | social | moral | financial | legal | epistemic | temporal
magnitude: negligible → existential
immediacy: distant → immediate
probability: speculative → certain
reversibility: easy → irreversible
visibility: private → public
```

Stakes must be demonstrated or explained; volume and dramatic music cannot manufacture them.

## 8. Character and speaker grammar

A narrative entity can perform one or more roles:

```text
PROTAGONIST
ANTAGONISTIC_FORCE
GUIDE
WITNESS
EXPERT
SKEPTIC
ALLY
FOIL
BENEFICIARY
DECISION_MAKER
NARRATOR
AUDIENCE_PROXY
SYSTEM_OR_INSTITUTION
```

For each significant entity, track:

```text
want
need
belief
misbelief
fear
constraint
power
knowledge
relationship
strategy
change
contradiction
voice
```

For nonfiction, “character arc” may mean a change in knowledge, position, method, relationship, or credibility—not a fabricated emotional transformation.

## 9. Point-of-view dimensions

```text
person: first | second | third
knowledge: omniscient | limited | witness | retrospective | real-time
reliability: reliable | qualified | contested | unreliable
access: external | internal | mixed
temporal position: before | during | after
stance: neutral | investigative | advocacy | personal | satirical | poetic
```

Decision rule:

```text
IF the central value is lived experience
THEN privilege first-person testimony.

IF the central value is verification
THEN privilege source visibility and an investigative narrator.

IF sources disagree
THEN represent disagreement in the narrative graph rather than smoothing it away.
```

## 10. Scene grammar

```text
SCENE
= place/time continuity
+ active subject
+ immediate objective
+ obstacle or uncertainty
+ behavior
+ turn
+ exit value
```

Scene value pairs may include:

```text
safe → unsafe
ignorant → informed
control → loss of control
trust → distrust
distance → intimacy
hope → disappointment
confusion → clarity
weakness → leverage
order → disorder
```

A scene without a turn may still function as atmosphere, observation, ritual, proof, or transition; label it accordingly.

## 11. Sequence grammar

A sequence combines beats or scenes around one local movement:

```text
SETUP → DEVELOPMENT → TURN → CONSEQUENCE
QUESTION → SEARCH → EVIDENCE → ANSWER
GOAL → ATTEMPTS → FAILURE → ADAPTATION
CLAIM → EXAMPLE → OBJECTION → QUALIFIED_SYNTHESIS
```

A complete episode or film combines sequences into a larger state change.

## 12. Reusable narrative architectures

### Explanation

```text
confusing observation
→ useful question
→ definitions
→ mechanism
→ example
→ boundary/exception
→ implication
→ usable model
```

### Argument

```text
claim
→ stakes
→ reasons
→ evidence
→ strongest counterargument
→ rebuttal or concession
→ qualified conclusion
→ action
```

### Investigation

```text
anomaly
→ initial hypothesis
→ source trail
→ conflicting evidence
→ failed explanation
→ new connection
→ best-supported finding
→ uncertainty ledger
```

### Transformation

```text
before state
→ dissatisfaction or inciting event
→ commitment
→ struggle
→ adaptation
→ threshold
→ result
→ cost and meaning
```

### Case study

```text
subject and baseline
→ objective
→ constraints
→ intervention
→ process
→ measurable outcome
→ alternative explanations
→ transferable lesson
```

### Tutorial

```text
outcome preview
→ prerequisites
→ model
→ ordered actions
→ checkpoints
→ common failures
→ verification
→ next application
```

### Interview or podcast

```text
identity/value promise
→ origin or context
→ central problem
→ mechanisms and stories
→ tension/objection
→ synthesis
→ personal meaning
→ practical takeaway
```

### Documentary portrait

```text
revealing present-tense behavior
→ context
→ desire
→ pressure
→ contradiction
→ consequence
→ reflective meaning
```

### Sales narrative

```text
recognized problem
→ cost of inaction
→ desired state
→ mechanism
→ proof
→ risk reversal
→ fit/qualification
→ action
```

### Horror or dread

```text
normal rule
→ subtle violation
→ denial
→ recurrence
→ pattern recognition
→ loss of control
→ confrontation
→ residue
```

Never use a genre pattern to falsify a nonfiction claim.

## 13. Hook grammar

A hook is a compact change in attention, not merely loudness.

```text
HOOK
= subject clarity
+ consequential gap
+ specificity
+ credible promise
+ forward vector
```

Hook operators:

```text
QUESTION: “Why did X happen?”
CONTRADICTION: “Everyone remembers X, but the first record says Y.”
OUTCOME_FIRST: show result before mechanism.
DEADLINE: establish time pressure.
IRREVERSIBLE_MOMENT: begin at the point of no return.
SPECIFIC_CONFESSION: precise personal disclosure with context.
VISIBLE_PROBLEM: demonstrate failure before explanation.
PATTERN_BREAK: violate an established audiovisual expectation.
```

Hook integrity test:

```text
Does the body answer the promised question?
Does the opening preserve the speaker’s meaning?
Is the uncertainty real?
Is the consequence proportional?
```

## 14. Pacing as a vector

Pacing is not simply “fast” or “slow.” Track:

```text
shot duration
beat duration
speech rate
idea density
new-entity rate
visual change rate
auditory change rate
question-to-answer gap
silence ratio
emotional change rate
tension slope
repetition interval
novelty interval
transition duration
```

### Pacing rules

```text
IF cognitive load rises
THEN reduce concurrent visual change OR slow speech OR add a summary.

IF emotion is the dominant value
THEN protect pauses, reactions, and unbroken performance.

IF procedural clarity is the dominant value
THEN align each action with one visible step and one verification cue.

IF energy remains flat across multiple beats
THEN vary scale, mode, question, evidence, rhythm, or emotional value—not all at once.
```

## 15. Emotional contour

Represent emotion per beat as a vector rather than one label:

```text
valence: negative ↔ positive
arousal: calm ↔ activated
control: powerless ↔ powerful
certainty: uncertain ↔ certain
connection: isolated ↔ connected
```

Create a target contour, then compare it with the observed performance. Do not force music or cutting to contradict authentic testimony without an explicit editorial reason.

## 16. Narrative compression

Compression removes time while preserving causal and ethical integrity.

Compression operators:

```text
remove repetition
merge equivalent propositions
replace examples with representative example
summarize process
use montage
use time card
use graphic model
use ellipsis
use narration bridge
use reaction instead of repeated explanation
```

Compression safety rules:

```text
KEEP qualifiers that change truth conditions.
KEEP causal links needed to understand responsibility.
KEEP uncertainty around disputed facts.
KEEP emotional pauses when the pause carries meaning.
KEEP enough chronology to prevent false implication.
```

## 17. Setup, callback, and payoff graph

```text
SETUP --prepares--> PAYOFF
MOTIF --repeats_with_change--> CALLBACK
QUESTION --resolved_by--> ANSWER
PREDICTION --confirmed_or_denied_by--> RESULT
PROMISE --fulfilled_by--> DEMONSTRATION
```

Payoff score:

```text
PayoffScore
= relevance to setup
× consequence
× surprise-with-coherence
× emotional or intellectual resolution
× recall strength
```

## 18. Narrative candidate generation with LAKA

For a base story graph, generate versions by change level.

### Baseline

```text
clean chronology
remove repetition
preserve speaker order
add minimal orientation
```

### Minor change

```text
alternate hook
move one example
shorten a section
add one visual explanation
change local pacing
```

### Major change

```text
replace chronological opening with outcome-first opening
center a different conflict
use a thematic rather than chronological sequence
change narrator stance
```

### Structural change

```text
intercut two timelines
build around question/evidence modules
convert interview into narrated documentary
convert long episode into chaptered argument
```

### Paradigm change

```text
change protagonist or audience proxy
turn explanation into investigation
turn interview into interactive branching narrative
make the process—not the result—the central story
```

Score candidates for:

```text
truth preservation
causal clarity
promise/payoff integrity
audience relevance
emotional coherence
novelty
retention potential
production feasibility
accessibility
risk
```

## 19. Narrative failure modes

```text
TOPIC_WITHOUT_CHANGE
CHRONOLOGY_WITHOUT_CAUSALITY
HOOK_WITHOUT_PAYOFF
CONFLICT_WITHOUT_STAKES
CLAIM_WITHOUT_EVIDENCE
EMOTION_WITHOUT_CONTEXT
MONTAGE_WITHOUT_ORIENTATION
EXCESSIVE_OPEN_LOOPS
FALSE_CAUSAL_IMPLICATION
OVER_COMPRESSION
REPETITION_WITHOUT_ESCALATION
SINGLE_EMOTIONAL_NOTE
UNMOTIVATED_TURN
ENDING_WITHOUT_NEW_STATE
CTA_WITHOUT_EARNED_RELEVANCE
```

## 20. Narrative acceptance test

A narrative passes when a reviewer can answer:

1. What changed?
2. For whom or what?
3. Why did it change?
4. What resisted the change?
5. Why does it matter to this audience?
6. What evidence supports the central claims?
7. Which important uncertainties remain?
8. What was promised, and where was it paid off?
9. What should the audience now know, feel, or do?
10. Could any edit create a materially false implication?
