Humaneness PerformanceNetAll references · Download PDF ↗
ContentsThe source becomes a virtual performanceThe shared taxonomy architectureRead the source, then make an actorCondition models, encode scenes and design examplesCharacter MakeupThe character interviewDramatic ObligationsActing Beats and Experience ChunksThe generation contractMara Solis / control and careTheo Marin / humor and connectionEvelyn Hart / evidence and admissionBuild, generate and reviewResearch foundationsVocal vocabulary and testingSources and implementation files

Script Analysis System

Humaneness
Script Analysis
System

Source material becomes a virtual actor’s performance design

Turn a book, script, screenplay, stage play or radio text into a persistent fictional character, the meaning of each scene, and ordered Acting Beats. Every beat contains concrete outward direction and exactly one authored Experience Chunk.

01

Character Makeup

The role across scenes.

02

Dramatic Obligations

All seven categories, including Emotional.

03

Beat + Experience

One outer specification and one inner chunk.

Part of Humaneness PerformanceNet
Generation specification · 2.0 · Formerly Actor-State Atlas

Use the companion Performance Annotation System to encode recorded real, reference or synthetic performances and to review a generated result.

Start here

The source becomes a virtual performance

The Humaneness Script Analysis System is the authoring and scene-encoding component of Humaneness PerformanceNet. It preserves why a fictional character acts, what a scene requires and how a local response should become speech or visible behavior.

Start with dramatic material: a novel, script, screenplay, stage play, radio play or an original authored scene. Read for facts, relationships, actions and contradictions. Build a Character Makeup for each fictional character. For every scene, formulate all seven Dramatic Obligations through that character’s particular makeup, including the scene’s Emotional obligation. Finally, arrange what is said, done and withheld into ordered Acting Beats.

One beat, one nested chunk. The outer Acting Beat carries exact dialogue, concrete action, silence, timing and delivery cues. Its single Experience Chunk carries the authored inner stream: attention, sensations, memories, associations, thoughts, emotion, competing impulses and continuity from the previous beat.

These inputs condition a generator. Audio output realizes the spoken line, breathing, rhythm and pauses. Audiovisual output can also realize gaze, face, body and action. The illustrated proposal also suggests keypoint sequences and avatar controls; these are intended uses of the representation, not demonstrated results.

“Experience” names fictional conditioning. A chunk describes a possible inner stream for the character. It does not assert that an AI model is conscious, that it feels the supplied emotion, or that a viewer can recover that inner stream from the output. A fear of loss can condition composed reassurance; the fear need not become directly detectable.

The division of work. This system stores source-grounded and authored character conditions. The Humaneness Performance Annotation System encodes observable real, reference and synthetic recordings, with interpretation and task-fit review separated. Both use shared taxonomy fields where appropriate. An authored condition is not converted into an observed fact merely because the field name is shared.

The three connected layers

The three connected layers
Overview supplied with the illustrated short guide. Read left to right as a conditioning workflow. The inner “experience” is authored fiction; the outputs shown are proposed generation targets. The selectable text before and after this plate gives the operational definition. Historical title in the supplied artwork: Actor-State Atlas; current system name: Humaneness Script Analysis System.

Humaneness PerformanceNet

Two systems, one shared descriptive space

Script Analysis SystemSource → character → sceneOrdered beats + one inner chunk eachAuthored conditions for virtual actorsPerformance Annotation SystemReal · reference · synthetic recordingsEvidence → description → task-fit reviewSeparate observation and interpretationShared taxonomies; different roles for each valuePerformanceNet · VoiceNet · EmoNet · situation and scenario vocabulariesCondition models · construct learning targets · select and diversify dataA taxonomy is a representation, not a trained model or an automatic reward.
Humaneness PerformanceNet joins two complementary systems through a shared descriptive space. A value’s source and role remain explicit when it is reused.
Shared componentWhat it contributesHow to use it
PerformanceNet Taxonomy82 descriptive fields: 41 inherited fields in Appendix A plus 41 additions in Appendix B. There are 79 ordinal scales and 3 categorical capture fields.Represent available performance attributes with exact anchors, source IDs, modality and missingness.
Observer-impression layer8 further scales in Appendix B, describing a particular viewer’s response.Keep viewer response and task-fit evaluation separate from the 82-field descriptive profile.
VoiceNetDetailed vocal qualities such as tempo, articulation, pitch range, respiration and texture.Pin the codebook. Map an authored voice direction or observed vocal description into compatible voice fields; do not equate similar names across taxonomies.
EmoNet40 emotion and state labels, with the selected intensity and expression conventions.Declare whether a label names authored inner state, intended display, observed expressive impression or a hypothesis.
Situation and scenario spaceThe shared Situation Taxonomy, plus situation, compound edge-case and extreme-premise source catalogs.Use the tags and premises to encode context, build conditions and inspect coverage. Catalog entries may overlap; their counts are not a sum of independent psychological states.

The Script Analysis System uses descriptors as conditions: a voice quality to realize, a situation to inhabit, a displayed affect to maintain. The Performance Annotation System uses supported descriptors to encode a recording. Source text, fictional Experience Chunks, observations and viewer responses stay distinct even when they refer to the same character and moment.

Read the source, then make an actor

Keep the text and the interpretation separable

Retain the original passage, scene or utterance and its locator. Mark which character facts are explicit in the source, which are interpretations of the dramatic material, and which are new directorial inventions. An adaptation can add a pause, a bodily action or a thought fragment; identify that addition instead of presenting it as the author’s wording.

MaterialWhat to extractWhat to design explicitly
Book or narrated proseActions, descriptions, relationships, narrated thoughts and changes over time.Which narration becomes dialogue, silent action or an Experience Chunk.
Screenplay or stage playDialogue, scene setting, stage directions, entrances, exits and dramatic reversals.The character-specific meaning of each scene and the local response between lines.
Radio play or speech scriptWords, speaker turns, audible actions, sound cues and silence.Breath, timing, delivery and any inner fragments that should condition them.
Original sceneThe declared premise, participants, text and intended constraints.All fictional background and inner material, marked as authored design.

Build one persistent character, then vary the encounter

Do not invent a new personality for every emotion label. The same character can command, comfort, joke or confess. A persistent makeup explains what remains recognisable and what changes when a different partner, risk or bodily circumstance becomes important. Licensed reference audio or video can supplement the written design across several behavioral modes; it does not replace the four portraits.

For a reference clip, store the locator, permission basis, behavioral mode and intended use. Say whether it illustrates voice identity, habitual movement, a delivery mode or a specific action. The examples in this volume contain no reference recordings, so their reference-media arrays are empty.

What a structured scene makes possible

Condition speech and audiovisual generators

Give an audio, video or audiovisual generator the source passage, persistent makeup, the current scene’s seven obligations and its ordered beats. Keep the exact spoken words in the outer beat and the unspoken fictional stream in its one Experience Chunk. A model can receive selected voice, expression, action and timing cues without losing the dramatic cause behind them.

Encode scenes for omni models

A structured scene can connect text, source context, permitted reference media and intended action in one representation. This allows a model input to distinguish who the character is, what the encounter means, what happens next and which cues belong to which modality. It is a proposed input organization, not a claim that a particular omni model already understands or reliably executes every field.

Design controlled examples and diverse coverage

Hold the words and character stable while changing one scene obligation, displayed affect, competing impulse or local cue. Use the Situation Taxonomy and scenario catalogs to vary partners, tasks, social constraints and physical circumstances. Record the condition that changed and the family of related variants so later filtering, evaluation or data splits can preserve the comparison.

Prepare the intended side of learning examples

The authored package can supply the requested conditions paired with a rendered take. The Annotation System supplies the separately supported description of that take. A reviewed training example can then pair a condition with eligible observed targets, or compare task-fit labels between alternatives. The script’s hidden feelings are conditioning fiction; they are not automatically correct output annotations or rewards.

The taxonomy supplies a language. A generator, supervised learner or reward model is a separate component that may consume selected representations. Neither the taxonomy nor a well-formed scene package is itself a trained model or a validated universal measure of acting quality.

Character Makeup

Write a persistent fictional role that the generation model can carry across scenes. Physical, emotional, intellectual and psychological makeup remain four distinct, connected portraits.

For an audiovisual actor, add authored appearance, habitual posture, movement and interaction with the setting. Connect the vocal consequences below to that broader embodiment. Appearance is a design choice from the source or adaptation, not a private fact inferred from a reference voice.

A descriptive profile is not a row of adjectives. It is an account of tendencies, causes, contradictions, contexts, and vocal consequences.

How to write each domain

Each of the four domains uses the same narrative grammar: default tendency → learned or imagined cause → condition of activation → contradictory pull → interpersonal consequence → audible implication. This makes a portrait specific enough for an actor, yet also usable as a generation condition. Scores can complement the portrait, but never replace it.

PART OF PORTRAIT GOOD QUESTION WHY IT CHANGES THE VOICE
Default tendency What does the person habitually do when nothing forces a choice? Sets the prior for rhythm, resonance, breath and interpersonal distance.
Conditional trigger Which situation reverses or exaggerates that habit? Identifies a learnable transition rather than a static style label.
Contradiction Which two needs can be active at the same time? Produces masked emotion, hesitation, contradictory prosody.
Relational variation How does behavior differ with authority, strangers, intimates? Allows the same voice identity to shift by partner.
Voice consequence What audible detail might carry the conflict? Links latent dramaturgy to acoustic affordances.
Epistemic status Who knows this and from what material? Prevents invented biography from passing as an audio measurement.

The four domains - specification level

Physical makeup

Stable and recurrent embodiment relevant to speech: habitual effort, carriage, breath placement, sensory habits, physical energy distribution, and recovery after strain. Do not infer bodily appearance, health diagnoses or exact age from audio alone.

Illustrative causal claim: A practiced crisis worker can retain clean articulation while involuntary breathing narrows under stress.

Emotional makeup

The character’s recurring emotional point of view, attachment habits, trigger sensitivity, emotional availability, and learned regulation style. Distinguish the experience of an emotion from its deliberately displayed mask.

Illustrative causal claim: A person who hides need under usefulness can sound most frightened at the moment they become exceptionally competent.

Intellectual makeup

Modes of attending, interpreting, reasoning and phrasing: analytical versus associative habits, verbal planning, tolerance of ambiguity, certainty calibration, and changes in cognition under load. This is NOT an IQ verdict.

Illustrative causal claim: A careful archivist may articulate evidence beautifully in public yet lose sentence structure when personally implicated.

Psychological makeup

Self-concept, conflicts, values, defensive habits, perceived safety, inner roles and long-range motivations. The domain is a fictional character design, not a diagnosis or a claim that private motives are audible facts.

Illustrative causal claim: A comic protector can sabotage intimacy by creating laughter just as quiet reciprocity becomes possible.

Character Makeup: the persistent role

Character Makeup: the persistent role
The four domains, fictional history and permitted reference media establish the character’s range. Descriptions of body, emotion and psychology are authored role material, not deductions about an AI system or a real person’s private traits. Historical title in the supplied artwork: Actor-State Atlas; current system name: Humaneness Script Analysis System.

A genuine Morris-inspired rehearsal bridge

The Character Obligation entry proposes comparing actor and role, imaging an antecedent life, and choosing stimuli that can generate equivalent impulses. Related entries discuss subpersonalities (available inner roles), externals (outside-in stimuli from people, objects and animals), and imaging (sensory-rich imagined circumstances). Translate these into optional director-provided preparation notes such as “locate a protective impulse,” not guessed truths about the actor. [02, 05-07]

Morris also names Emotional Point of View: a persistent feeling-tone through which a role perceives its world. It is useful as a slowly varying bridge between makeup and the individual scene’s emotional obligation. [13]

The character interview

Questions designed to elicit thick, playable prose rather than shallow categorical tokens.

Physical domain

  • 01. Where does the body hold effort when this person is trying to look unaffected?
  • 02. What does ordinary breath feel like before anyone makes a demand?
  • 03. What repetitive job, craft, habit, or environment has shaped their motor economy?
  • 04. How does their vocal behavior change after exertion, pain, fatigue, cold, or relief?
  • 05. Which physical gesture accompanies a lie, apology, confession, reassurance, or refusal?
  • 06. What can their body do easily, and what requires deliberate management?
  • 07. What is their instinctive interpersonal distance and projection level?
  • 08. Which physical habit disappears around a beloved or feared partner?
  • 09. What precise audible consequence should this physical design have - and which should it NOT imply?

Emotional domain

  • 01. What emotional weather does the character carry into a neutral room?
  • 02. Which emotion is easiest to show? Which is hardest to allow another person to see?
  • 03. What triggers tenderness, envy, shame, anger, grief, relief, or play?
  • 04. What is felt, what is displayed, and who benefits from the difference?
  • 05. What emotion is a second-order response to another emotion (anger over fear, humor over shame)?
  • 06. How quickly do feelings rise, decay, recur or cross-contaminate?
  • 07. When does the character solicit comfort directly, indirectly, or not at all?
  • 08. Which partner makes them emotionally younger, older, braver or smaller?
  • 09. What will happen to the voice when emotional regulation succeeds - and when it slips?

Intellectual domain

  • 01. Does thinking proceed in proofs, procedures, images, associations, narratives, or questions?
  • 02. How much verbal planning takes place before speech begins?
  • 03. How does the person distinguish fact, interpretation, hope and fear?
  • 04. What kind of ambiguity do they tolerate intellectually but not personally?
  • 05. When challenged, do they explain, inquire, attack, joke, avoid, or go quiet?
  • 06. Where does education or specialized craft show up in phrasing, not social stereotypes?
  • 07. Does time pressure sharpen thought, flatten it, or make it jump?
  • 08. What sound marks genuine discovery rather than an already rehearsed conclusion?
  • 09. What intellectual competence becomes a relational liability?

Psychological domain

  • 01. What must remain true for the character to feel worthy of love, respect or safety?
  • 02. What is the preferred self-image, and which behavior undermines it?
  • 03. What private fear is protected by their most admired public strength?
  • 04. Which opposing inner roles can take control when the context changes?
  • 05. How do they defend against shame, helplessness, dependency or abandonment?
  • 06. What moral rule is non-negotiable until a loved person is implicated?
  • 07. What recurring story do they tell about their own behavior?
  • 08. What does the character need but have difficulty requesting?
  • 09. What single event could force a revision of their self-image?

Answering standard: Give each domain at least one substantial paragraph. Every claim should have a behavioral mechanism, an enabling circumstance, and a potential audible signature. A profile made only of adjectives (“confident, kind, anxious”) is considered incomplete.

Dramatic Obligations

Dramatic Obligations describe what the current scene means to this particular character. They are interpreted through the Character Makeup, not attached as a generic situation label.

The Emotional obligation is required. State which feelings must be present, mixed, redirected or concealed in this scene. Keep it alongside Time & Place, Relationship, Character, Historic, Thematic and Subtextual. A scene goal or social role does not replace its emotional reality.

For each obligation, connect the source circumstance to a character-specific response. A hospital corridor may be physically quiet for everyone, but for Mara its quiet removes the protection of command language. Her need to be useful changes what an apology costs. The six examples below each include a short “Through this character’s makeup” bridge before all seven obligations.

The seven Morris categories are the foundation; new scene obligations add action, social regulation and model-facing controls.

The seven Morris obligations describe different realities a role must make playable. They are not seven interchangeable emotion labels. The period-specific Historic Obligation must not be collapsed into personal backstory: the latter is material for characterization, while the former concerns the customs and conditions of the historical world. [04, 12]

MORRIS OBLIGATION AUTHORING QUESTION
Time & Place What immediate space, time, soundscape, distance and physical arrangement affect the voice?
Relationship Who exactly are the other people, and how does the character feel toward each one?
Emotional Which feelings must be present, mixed, redirected or hidden in the scene?
Character Which established personal tendencies and contradictions must come alive here?
Historic Which period norms, technologies, manners, speech expectations or moral assumptions matter?
Thematic What larger dramatic human question does this encounter embody?
Subtextual What unsaid emotional reality persists beneath the literal words?

Dramatic Obligations: this character in this scene

Dramatic Obligations: this character in this scene
Use all seven categories: Time & Place, Relationship, Emotional, Character, Historic, Thematic and Subtextual. “Historical” and “Sub-textual” in the illustration name the same inherited categories. Explore their meaning through the fictional character’s makeup and felt response. Historical title in the supplied artwork: Actor-State Atlas; current system name: Humaneness Script Analysis System.

Relational obligation: a multi-partner matrix

The knowledge-base entry explicitly separates (1) who a partner is to the role and (2) how the role feels toward that partner. Preserve this distinction for every relevant participant, not just for the person being addressed at the instant of speech. A speaker can reassure a junior colleague, fear for an absent responder, and resent a superior - simultaneously. [11]

PARTNER FIELD QUESTION EXAMPLE
relation / history What shared history makes the person matter? Trusted junior colleague, sent into danger
felt_toward What contradictory emotions are directed at this partner? Protectiveness, guilt, frustration
desired_effect What response does the speaker seek from this partner? Answer the radio; breathe before reporting
power / norms Who can grant, deny, punish or abandon? Formal boss vs. emotional vulnerability
attention_now Is this person addressed, imagined, absent, overhearing? Missing responder, physically absent

Proposed obligations beyond Morris

NEW PROPOSED FIELD SCENE-LEVEL QUESTION
objective / tactic What is the speaker trying to cause in the listener, and how?
obstacle / stakes What resists the effort, and what happens if it fails?
social contract Which role, decorum, duty or face-saving demand regulates expression?
status / power Who controls resources, the floor, affection or institutional legitimacy?
epistemic gap What is known, believed, concealed, doubted or misread?
physiology / sensory Which current bodily or environmental constraints shape speech?
regulation / masking What is felt vs shown; how successfully is it regulated?
anticipation What reply or consequence is being hoped for or feared?

These extensions are a deliberate design choice. Playable actions and obstacles are informed by broader actor-training approaches, especially Uta Hagen’s questions and Stanislavsky-based scene analysis; they should not be attributed to Morris as additional canonical obligations. [19, 20, 22]

Acting Beats and Experience Chunks

A beat is a short, ordered unit of dramatic action. Give it one outward specification and exactly one nested Experience Chunk. The two describe the same moment at different levels.

The outer Acting Beat

Write the exact words and the concrete actions the model should realize. Include an intentional silence where it matters. Add relative order, approximate or prescribed timing when available, and usable voice or physical cues. “Conceal fear” is an inner task; “steady onset, shortened intake, wait before the name” gives the generator a possible realization. Avoid prescribing every muscle or syllable when a less rigid condition can leave room for natural variation.

The single inner chunk

The Experience Chunk is a coherent authored stream rather than a second independent timeline. It may contain sensory impressions, recalled or imagined material, a mundane association, an unfinished thought, an emotion and several competing urges at once. Keep one chunk object per beat even when its contents are mixed. Use empty arrays for dimensions that have no useful authored content; an elaborate memory is not required in every beat.

Choose a boundary because something changes

Begin a new beat when attention, information, impulse, tactic, relationship or bodily constraint changes. A pause or camera cut can coincide with that change, but neither automatically defines it. If two changes need different local directions, split them into two ordered beats, each with its own single chunk.

In the illustrated Juliet example, the authored stream can combine night air, an imagined Romeo, fear of the feud and an unrelated passing thought. That stream conditions the spoken and visible action. It is not a transcription of an AI model’s cognition, and the generated face or voice does not validate it as a hidden fact.

One Acting Beat, one Experience Chunk

One Acting Beat, one Experience Chunk
Concrete text, action, silence and delivery belong to the outer beat; attention, thought, sensation and impulses belong to its one nested chunk. The illustration’s “inner experience” is fictional conditioning. Audio, video, keypoints and avatar animation are proposed outputs, not results demonstrated by this document. Historical title in the supplied artwork: Actor-State Atlas; current system name: Humaneness Script Analysis System.

The inner stream: field guide

Inner fieldWhat to write
AttentionWhat the character notices now; selection matters as much as the stimulus itself.
SensationAn authored sensory impression or bodily feeling, distinguished from the outer cue that may express it.
Memory / imaginationA remembered image, imagined reply or anticipated consequence that is active now.
AssociationA connection the present moment calls up, including a stray or mundane one if it serves the portrayal.
Thought fragmentsUnspoken wording or incomplete ideas. These are conditioning, not additional spoken dialogue.
Emotion / impulseWhat the fictional character feels and is urged to do, including regulation and a competing urge.
CarryoverWhat remains active from the preceding beat and what the new event changes.
Reading the revised examples. The original dialogue, character portraits, scene obligations, stimulus/effect, impulses, displayed affect and vocal directions remain. Attention, thought fragments, selected sensations, memories, associations, competing impulses and carryover are explicit new design additions. Each beat now exposes one outer specification and one Experience Chunk; no new performance recordings or measured scores are claimed.

The generation contract

Pass the source, persistent character and current scene together. Keep the ordered beats under that scene, with each Experience Chunk nested directly in its own beat.

ObjectRequired relationPurpose
Source materialOne or more identified source textsPreserve the dramatic text and distinguish interpretation from invention.
CharacterFour makeup domains; optional permitted reference mediaKeep the role coherent across scenes and expressive modes.
SceneAll seven obligations interpreted through that makeupGive this encounter its particular meaning, constraints and emotional reality.
Ordered Acting BeatOne outer specification + exactly one Experience ChunkBind local outward behavior to one authored inner stream.
Output requestRequested modality and realization constraintsAsk for audio, video, audiovisual performance or a supported motion representation.

Do not confuse a request with a result

The outer beat describes desired behavior. The chunk describes fictional inner conditioning. Neither is an observation of a recording. Keep rendering identifiers, actual durations, transcripts and measurements in a separate media record linked by IDs. The Performance Annotation System can describe that output and collect task-fit judgements under the declared modality and context.

A valid beat object

The following is valid JSON for one beat. Empty arrays are intentional: this compact variant adds no sensory image, memory, association or physical action. The supplied full example package contains the richer chunks printed in the dossiers. Null times mean that no exact time was supplied; they do not mean zero seconds.

{
  "id": "mara-solis-scene-1-beat-01",
  "order": 1,
  "outer": {
    "text": "Keep the channel open.",
    "text_anchor": "Keep the channel open.",
    "playable_action": "stabilize",
    "displayed_affect": "authority",
    "action_cues": [],
    "silence_cue": null,
    "timing": {
      "start_s": null,
      "end_s": null,
      "direction": "Ordered relative to the preceding beat; exact duration is not supplied."
    },
    "vocal_cues": [
      "Dry, firm attack on Keep; moderate pace; no theatrical shout."
    ],
    "physical_cues": [],
    "voice_targets": {
      "TEMP": 3,
      "CLRT": 5,
      "TENS": 4,
      "EMPH": 4
    }
  },
  "experience_chunk": {
    "status": "authored_fiction",
    "stimulus_and_effect": "A gap of radio silence threatens the only link. Alarm focuses into a narrow operational priority.",
    "attention": "The silence on the radio and the fragile communication link.",
    "sensations": [],
    "memories": [],
    "associations": [],
    "thought_fragments": [
      "Keep this one connection alive."
    ],
    "emotion": "fear",
    "impulse_and_regulation": "Demand connection immediately.",
    "competing_impulse": "Call for reassurance about Eli before the communication task is secured.",
    "carryover": null
  },
  "source_emotion_labels": [
    "Fear",
    "Concentration"
  ]
}

Machine companions

actor-state-generation.schema.json defines the complete package. actor-state-generation.examples.json contains all three characters, six scenes and twenty-four beats, with the full portraits and scene obligations. These are newly authored companions to this edition and have been validated locally. They are separate from the unavailable files named in the earlier source PDF.

Human-readable control targets

The dossiers retain their original illustrative voice targets and display names beside the codes. A target such as Tempo (TEMP) = 2 is a requested codebook setting, not a measured rate or proof of an emotion. Use the exact selected codebook and inspect its anchors before generation. The prose direction remains the primary reading aid; a numeric target does not override an incompatible dramatic instruction.

Inherited emotion labels need a role. The original EmoNet candidate names are preserved in source_emotion_labels as metadata for the whole beat. Some concern the character’s authored inner emotion; others may concern its public display. They are not automatically displayed-affect controls. Before using a label as a control, explicitly decide which role it serves. For example, the first Mara beat retains “Fear” and “Concentration” from the source while its inner emotion is fear and its intended public display is authority. The operative fields remain experience_chunk.emotion and outer.displayed_affect.

TargetRead it as
Tempo TEMPspeech speed
Articulation Clarity CLRTclarity of consonants and vowels
Tension TENSperceived tightness in vocal delivery
Emphasis EMPHstress on selected words or syllables
Respiration RESPbreathing audibility and effort
Vulnerability VULNexpressed openness or guardedness
Pitch Range RANGwidth of pitch movement
Warmth WARMwarmth from low-mid-frequency sound
Disfluency DFLUfillers, false starts and self-corrections
Playful Style S_PLAYplayful, lighthearted delivery
Smoothness SMTHsmoothness and regularity of speech timing
Stance STNCsocial stance in the exchange

The legend follows the pinned source recorded in voice-code-legend.json. In particular, WARM describes perceived low-mid-frequency vocal warmth; it is not an interpersonal-warmth score. Vocal targets remain illustrative instructions, with codebook-specific 0–6 anchors.

Share fields without merging claims

The two systems can describe the same scene without claiming the same knowledge. Preserve the original text, authored design, media and observations as separate records. When a generated take or reference rendition corresponds to a script, link it explicitly; observed timing and segmentation may differ from the planned beats.

Interface fieldWhat it preserves
Source, scene and participant IDsThe material and focal person or character. Use a recording-local participant ID when no fictional role is known.
Planned beat / observed interval alignmentAn optional link between an authored beat and an observed time span. It need not be one-to-one and must not force a recording into the intended timing.
Field code + taxonomy versionThe exact descriptor and its interpretation; preserve ordinal, categorical and event types.
Role + basisAuthored condition, direct observation, supported context, inference, viewer response, task-fit label or derived training target.
Modality + availabilityAudio, video or audiovisual access; which regions/events can be rated, with null and a reason for unavailable fields.
Evidence + calibrationThe source span or media interval, rater/model provenance, uncertainty, alternatives and protocol revision.
Learning or selection useWhich labels are eligible for supervision, preference/reward learning, retrieval or sampling, and why.

Do not fill an unobserved inner state from a script and then label it “observed.” Do not convert a missing visual score into zero for an audio-only example. Do not use a high ordinal value as an automatic positive reward: the high end may describe stronger tension, hostility or another task-dependent attribute.

Mara Solis

Authored character design · two scenes · eight ordered beats. Every beat contains one Experience Chunk. Original portraits and scene detail are retained; new inner-fragment additions are fictional conditioning.

Control is her language of care. Silence is what she cannot manage.

FLOOD-CONTROL OPERATIONS COORDINATOR • CONTEMPORARY PORT CITY; EMERGENCY OPERATIONS AND FAMILY LIFE

Long-range identity: Control is her language of care. Silence is what she cannot manage.

Character Makeup / four-domain portrait

Physical disposition

Mara has learned to conserve movement in rooms where panic spreads faster than information. Her shoulders remain apparently level, but she locks the muscles beneath the collarbones when she is accountable for another person's safety. Years of radio work have given her a practical voice: breaths placed before instructions, consonants clean enough to survive static, and deliberate pauses that invite confirmation. She rarely looks outwardly exhausted; fatigue appears first as an extra swallow and a narrowing of her phrases. In private, her body takes much longer than her words to recognize that the emergency is over. This is a playable bodily design, not a claim about a medical condition.

Voice consequence. Low-amplitude starts, clean consonant attacks, economy of breath, and task-first phrasing; pressure tends to sharpen rather than amplify her voice.

Countertendency. After immediate danger passes, delayed bodily release can make even a small admission difficult to voice.

Emotional disposition

Mara feels alarm intensely but considers open alarm an additional burden she would be placing on others. Her habitual emotional point of view is vigilant responsibility: she notices who needs steadiness before she notices her own loneliness. Affection is expressed through useful acts, precise remembering, and preparedness, while personal requests for reassurance feel shamefully indulgent to her. Under urgent pressure she suppresses visible fear; under intimate pressure she suppresses need. Those are different forms of the same protective reflex. She can experience anger when people take risks, yet that anger is often secondary to an almost unbearable fear of losing them. Her tenderness becomes audible only when she ceases to organize the other person's response.

Voice consequence. Her fear often sounds like concentration; deep attachment arrives as a softened final consonant or a sentence that cannot be completed with a directive.

Countertendency. Suppressed fear can leak through breath timing even when pitch and loudness remain controlled.

Intellectual disposition

Mara thinks in dependencies: which gate is open, who has confirmed a level, which piece of evidence is still missing. She is not naturally unemotional; she has trained herself to translate strong feeling into an ordered series of verifiable actions. She distrusts grand declarations because they cannot be checked, and she interrogates the meaning of every ambiguous word before making a consequential choice. In technical conversation her reasoning is rapid and compressed; in a family conversation she can become paradoxically slow, searching for a form of truth that cannot be written as a procedure. The intellectual tension in her portrayal is between procedural competence and the irreducible uncertainty of human relationships.

Voice consequence. Clause structure is modular, numbers receive clear stress, and revisions sound intentional until an emotionally unsolvable question disrupts syntax.

Countertendency. Precision that is helpful professionally can be perceived as avoidance by loved ones.

Psychological disposition

Mara's organizing belief is that she earns belonging by being the person who does not fall apart. This belief makes her admirable and costly: others trust her under fire, but some stop asking whether she can bear what she carries. Her deepest conflict is not courage against cowardice; it is service against the wish to be looked after. She sees herself as honest because she almost never tells literal lies, while overlooking how withholding grief can become its own kind of deception. A useful inner counter-role is the frightened daughter she has long refused to let other people see. When challenged by someone she loves, her defense is to offer a solution before admitting harm.

Voice consequence. Defensiveness is often expressed as helpfulness; authentic vulnerability is the brief loss of the rehearsed answer, not necessarily tears.

Countertendency. She may resist forgiveness that comes before she has named precisely what she did.

Scene 1 / The flooded control room

01:42, COASTAL FLOOD-CONTROL CENTER, INTERMITTENT RADIO CONTACT DURING A GATE FAILURE

The flood gate has failed. A young field responder, Eli, is missing from the video feed; junior operator Nico is beginning to panic. Mara must keep the radio usable while the room turns toward her for certainty she does not possess.

“Keep the channel open. Eli, answer only if you can hear me. Nico, take one breath. Then tell me the water level.”

Through this character’s makeup. Mara’s trained economy of movement and procedural thinking turn fear into usable instructions. Her emotional habit is to protect others from her alarm; her need to earn belonging through competence makes Eli’s silence personally threatening.

The seven dramatic obligations

Time & Place. At 01:42, alarm tones, damp clothing, a wall of changing measurements, and a noisy radio tighten every available breath. The place is public, operational, and acoustically unforgiving.

Relationship. Nico is her anxious junior colleague, whom she protects without infantilizing; Eli is the responder she sent toward the gate. She needs both, but only Nico can currently hear her clearly.

Emotional. She feels fear for Eli and a surge of responsibility for Nico. She displays concentration and procedural calm; the gap between felt alarm and shown steadiness is the scene's central pressure.

Character. Her habit of reducing chaos to steps becomes useful and defensive at once. Her reluctance to request comfort keeps her tone functional even when a name becomes hard to speak.

Historic. Contemporary emergency culture values clear protocols, recorded decisions, radio discipline, and accountability. These conventions are social and professional, not Mara's personal backstory.

Thematic. Can responsibility protect other people without turning into a refusal to be vulnerable? The scene frames competence as both gift and prison.

Subtextual. Beneath the instructions: 'Please still be alive, Eli; please let Nico believe I know what comes next.' The unsaid fear rides underneath ordinary operational language.

Proposed extension obligations

Playable objective. Keep both partners responsive long enough to obtain a reliable water-level reading. Stakes. Someone may be injured if the room loses coordination or acts on unconfirmed information.

Obstacles. Radio interference; Missing visual confirmation; Nico’s rising panic; Mara’s own anticipatory guilt.

Social / status. Leader is expected to remain legible and composed; personal panic would consume shared bandwidth. High formal authority over Nico, little immediate control over the field situation.

Knowledge / regulation. She cannot verify Eli’s condition; she explicitly chooses not to claim that she can. Suppress audible panic without deadening the task-based urgency.

Current physiology. Held upper-chest tension, controlled intake before commands, mild breath shortening as uncertainty grows.

LAION situation dimensions for later codebook matching: SOCT · TRGT · ENVI · FATG · POSE

Partner-specific obligations

Junior Operator / nico. Teacher-like trust mixed with fear of failing him; she uses simple steps rather than reprimand. Desired response: Nico regains agency and reads the level.

Field Responder / eli. Professional responsibility with protectiveness she has not admitted to herself. Desired response: Eli gives any intelligible sign of contact.

Ordered Acting Beats

BEAT 01 · stabilize

“Keep the channel open.”

Outer Acting Beat. Dry, firm attack on Keep; moderate pace; no theatrical shout.

Displayed affect. authority.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. A gap of radio silence threatens the only link. Alarm focuses into a narrow operational priority.

Attention / thought. The silence on the radio and the fragile communication link. Unspoken fragment: “Keep this one connection alive.”

Sensation. Tightness high in the chest. Association. An open channel means a remaining chance of contact.

Emotion. fear. Impulse and regulation. Demand connection immediately. Competing impulse. Call for reassurance about Eli before the communication task is secured.

Carryover. Begin from the character’s scene obligations; no earlier beat in this scene is supplied.

Vocal targets. Tempo (TEMP) = 3; Articulation Clarity (CLRT) = 5; Tension (TENS) = 4; Emphasis (EMPH) = 4. Inherited whole-beat emotion labels (unassigned): Fear, Concentration.

BEAT 02 · reach

“Eli, answer only if you can hear me.”

Outer Acting Beat. Name is warmer and marginally longer; breath almost catches before answer.

Displayed affect. controlled concern.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. No visible evidence confirms Eli is safe. His name briefly pierces her professional distance.

Attention / thought. Eli’s name and the absence of confirmation. Unspoken fragment: “Please be there.”

Sensation. A momentary catch before the next word. Memory. She sent Eli toward the gate. Association. A name briefly becomes a person rather than an operational role.

Emotion. fear mixed with attachment. Impulse and regulation. Ask whether he is alive; stop herself from pleading. Competing impulse. Plead for proof that he is alive.

Carryover. Carry forward the earlier fear; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 2; Respiration (RESP) = 4; Vulnerability (VULN) = 2; Pitch Range (RANG) = 2. Inherited whole-beat emotion labels (unassigned): Fear, Longing.

BEAT 03 · regulate partner

“Nico, take one breath.”

Outer Acting Beat. One breath is paced with space around it; consonants soften while control remains.

Displayed affect. reassurance.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Nico begins speaking too quickly to report. Her attention switches from missing data to the person in front of her.

Attention / thought. Nico’s hurried reporting. Unspoken fragment: “One thing at a time; stay with me.”

Association. The task needs a person who can breathe and answer.

Emotion. protective alarm. Impulse and regulation. Borrow calm for him. Competing impulse. Rush him because the missing number feels urgent.

Carryover. Carry forward the earlier fear mixed with attachment; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 2; Warmth (WARM) = 4; Emphasis (EMPH) = 3. Inherited whole-beat emotion labels (unassigned): Concentration.

BEAT 04 · direct

“Then tell me the water level.”

Outer Acting Beat. Pitch settles, final phrase becomes brisk, and level takes stress.

Displayed affect. task focus.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. There is now a chance to recover usable information. Relief cannot settle until a number arrives.

Attention / thought. The possibility of obtaining a reliable water level. Unspoken fragment: “Now give me something we can act on.”

Association. A number gives uncertainty a next action.

Emotion. urgent hope. Impulse and regulation. Convert fragile steadiness into action. Competing impulse. Settle into relief before the result is known.

Carryover. Carry forward the earlier protective alarm; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 4; Articulation Clarity (CLRT) = 5; Emphasis (EMPH) = 5. Inherited whole-beat emotion labels (unassigned): Concentration.

Why this scene matters for controllable speech generation. The identical person’s stable habits are maintained while her relational availability changes. The voice arc should not collapse into a single emotion for the whole clip.

Scene 2 / The call she did not make

TWO DAYS LATER, FLUORESCENT HOSPITAL CORRIDOR AFTER THE IMMEDIATE CRISIS

Mara’s adult daughter Lena has learned that her mother was safe enough to send updates to colleagues but did not respond to a personal message. The disaster is over; Mara now faces a truth no procedure can rescue her from.

“I did hear you. I kept thinking there would be a better moment to call. There wasn't. I'm sorry.”

Through this character’s makeup. The same controlled body and orderly reasoning now create distance from Lena. Mara’s care remains real, but her habit of offering a solution must give way to admitting a need and accepting an answer she cannot manage.

The seven dramatic obligations

Time & Place. Late afternoon in a corridor that is quiet in the wrong way: vending-machine hum replaces alarms. Nothing outside requires command language, so every clipped instruction sounds conspicuous.

Relationship. Lena is Mara's daughter and most persistent witness. She admires Mara's bravery but resents having to guess when the professional role ends and her mother begins.

Emotional. Mara feels shame, affection, and apprehension; she initially offers an explanation that resembles an incident report. The scene asks her to permit regret to replace competence as her audible center.

Character. The same orderly mind that saved time in the flood room now creates emotional distance. Her physical composure becomes a visible attempt not to fall apart in front of her daughter.

Historic. Contemporary parent-child expectations include constant digital availability, making silence feel chosen even when produced by overload and fear.

Thematic. Being indispensable to strangers does not excuse becoming inaccessible to those closest to you.

Subtextual. Under 'I did hear you' lies 'I was frightened of needing you.' Under 'I'm sorry' lies an unfamiliar request: 'Please stay with me while I cannot fix this.'

Proposed extension obligations

Playable objective. Tell the truth without forcing Lena to absolve her. Stakes. Their closeness may remain conditional on Mara’s usefulness rather than mutual trust.

Obstacles. Habitual self-justification; Fear of appearing weak; Lena’s justified anger.

Social / status. Unlike the control room, authority buys her nothing; Lena wants an accountable parent. An old parental authority is present, but Lena controls whether the conversation continues.

Knowledge / regulation. Mara can describe her behavior but cannot know how Lena will receive the apology. Let emotion emerge without using it to solicit instant forgiveness.

Current physiology. Delayed exhaustion and a jaw that releases only after she abandons the explanation.

LAION situation dimensions for later codebook matching: SOCT · TRGT · FATG · ENVI

Partner-specific obligations

Adult Daughter / lena. Love complicated by repeated emotional absence; status shifts from commanding mother to accountable partner. Desired response: Lena hears an apology, not a defense.

Ordered Acting Beats

BEAT 01 · clarify

“I did hear you.”

Outer Acting Beat. Precise, slightly over-articulated did, more defensive than intended.

Displayed affect. controlled honesty.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Lena accuses her of ignoring the message. A reflex to defend factual accuracy rises.

Attention / thought. Lena’s accusation that the message was ignored. Unspoken fragment: “That is not exactly what happened.”

Sensation. Residual bodily fatigue after the crisis. Memory. The message arrived; Mara did hear it. Association. Factual correction offers the familiar shelter of a report.

Emotion. shame. Impulse and regulation. Correct the record before acknowledging the wound. Competing impulse. Acknowledge the injury before defending the literal record.

Carryover. Begin from the character’s scene obligations; no earlier beat in this scene is supplied.

Vocal targets. Articulation Clarity (CLRT) = 5; Tension (TENS) = 4; Emphasis (EMPH) = 4. Inherited whole-beat emotion labels (unassigned): Concentration.

BEAT 02 · confess

“I kept thinking there would be a better moment to call.”

Outer Acting Beat. The sentence lengthens and thins; a tiny hesitation before better.

Displayed affect. uncertain openness.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Her own explanation sounds inadequate as she says it. Reasoning dissolves into awareness of avoidance.

Attention / thought. The inadequacy of her own explanation as she hears it. Unspoken fragment: “I kept making it later.”

Memory. The call repeatedly postponed. Association. A better moment begins to sound like an excuse.

Emotion. regret. Impulse and regulation. Keep explaining, then let the thought lose momentum. Competing impulse. Continue explaining until the delay seems reasonable.

Carryover. Carry forward the earlier shame; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 2; Disfluency (DFLU) = 3; Vulnerability (VULN) = 4. Inherited whole-beat emotion labels (unassigned): Sadness.

BEAT 03 · concede

“There wasn't.”

Outer Acting Beat. Short and unadorned, ending downward without a performative sob.

Displayed affect. plain regret.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. She realizes no future moment can retroactively repair the silence. A small acceptance arrives.

Attention / thought. The impossibility of repairing the earlier silence retroactively. Unspoken fragment: “There was no better moment.”

Memory. There was no later call that could erase the omission.

Emotion. grief. Impulse and regulation. Stop defending. Competing impulse. Add one more qualification to soften responsibility.

Carryover. Carry forward the earlier regret; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 2; Emphasis (EMPH) = 3; Pitch Range (RANG) = 1. Inherited whole-beat emotion labels (unassigned): Sadness.

BEAT 04 · apologize

“I'm sorry.”

Outer Acting Beat. Softer onset, breath exposed, no insistence on the listener’s response.

Displayed affect. unmasked care.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Lena has not turned away. Love becomes a risk rather than a duty.

Attention / thought. Lena has remained in the conversation. Unspoken fragment: “Let the apology stand.”

Sensation. Breath becomes less guarded. Association. Being accompanied may require admitting a need.

Emotion. love and vulnerability. Impulse and regulation. Offer apology and leave room for her answer. Competing impulse. Seek immediate forgiveness or organise Lena’s answer.

Carryover. Carry forward the earlier grief; the new stimulus changes its expression.

Vocal targets. Warmth (WARM) = 5; Vulnerability (VULN) = 5; Tension (TENS) = 2. Inherited whole-beat emotion labels (unassigned): Sadness, Longing.

Why this scene matters for controllable speech generation. The identical person’s stable habits are maintained while her relational availability changes. The voice arc should not collapse into a single emotion for the whole clip.

Theo Marin

Authored character design · two scenes · eight ordered beats. Every beat contains one Experience Chunk. Original portraits and scene detail are retained; new inner-fragment additions are fictional conditioning.

He jokes to keep the door open, then fears anyone walking through it.

COMMUNITY ARTS FACILITATOR AND OCCASIONAL STAND-UP PERFORMER • CONTEMPORARY REGIONAL TOWN, TRANSIT PLATFORMS AND A NEIGHBORHOOD ARTS HALL

Long-range identity: He jokes to keep the door open, then fears anyone walking through it.

Character Makeup / four-domain portrait

Physical disposition

Theo’s body enters conversations before his conclusions do. He tilts toward people, organizes attention with his hands, and uses tiny musical changes of pace to tell an audience where to look. His quick respiratory recovery makes jokes feel effortless, but when an encounter becomes personally consequential, his breath stops arriving at the same convenient points. He habitually smiles through vowels, creating bright resonance that can be mistaken for emotional safety. Alone, he folds inward and grows still, not because he has a secret illness but because performance has temporarily exhausted his preferred way of managing contact. Physical exuberance is his social instrument, not his entire identity.

Voice consequence. Bright inflection, unexpected syllabic speed-ups, elastic pauses before punch lines, and an inviting forward focus.

Countertendency. When sincerity displaces the joke, the sudden absence of bounce reveals more than a loud confession would.

Emotional disposition

Theo genuinely delights in other people, especially when he can give them relief from an awkward or heavy moment. His comic spontaneity is therefore both authentic play and strategic concealment. Anger at abandonment is hard for him to own directly; it emerges first as exaggerated teasing and only later as loneliness. He is quick to detect rejection and even quicker to signal that he has not been injured by it. He longs for stable attachment but resists any conversation in which he must ask for it without the protection of a joke. Unlike a uniformly 'happy' character, his emotional baseline changes with perceived availability: full of light when connection seems possible, abruptly watchful when the other person hesitates.

Voice consequence. Playfulness can mask grief without eliminating it; a punch line may break its own rhythm when the hidden emotion leaks.

Countertendency. Sincerity and comedy sometimes coexist inside the same phrase, creating mixed signals rather than a clean mood switch.

Intellectual disposition

Theo thinks associatively and in social patterns. He remembers the exact phrase that made someone laugh months ago and can reframe a hostile remark before anyone else has processed it. He is perceptive about mood but unreliable in his predictions of what other people privately intend; he often interprets silence as rejection when it may be thoughtfulness. He is verbally inventive and uses metaphors faster than he can test their truth. His conflict is that improvisational intelligence helps him build immediate rapport but can keep him from completing a single frighteningly direct sentence. Under relational strain, a cascade of clever ideas becomes a sophisticated form of interruption.

Voice consequence. Fast association, colorful stress patterns, elastic timing; genuine discovery can be heard when word search replaces prepared cleverness.

Countertendency. He mistakes quick verbal repair for genuine resolution of the other person's hurt.

Psychological disposition

Theo's private rule is that people stay if being with him feels easy. He calls this generosity; sometimes it is a quiet bargain in which he purchases company with entertainment. He has an admirable capacity to turn shame into shared amusement, yet he fears the moment after laughter when the other person can still choose to leave. A disowned part of him wants to be boring, tired, angry, or needy without losing love. He initially treats intimacy as a performance problem and only slowly accepts that another person may wish to remain when there is no performance to reward. The useful acting tension is between his impulse to charm and the harder impulse to remain present.

Voice consequence. He tends to use jokes as deflections; his most truthful moments contain unfilled silences and simple diction.

Countertendency. He may sabotage a tender exchange with one final joke just as emotional safety becomes available.

Scene 1 / Platform four

23:11, NEAR-EMPTY TRAIN PLATFORM; VENDING MACHINES AND DISTANT ARRIVALS

Theo's older sister Ada has returned after nearly a year of scarce contact. He wants an explanation and reassurance but opens with a bit of stand-up, testing whether she will laugh rather than asking why she disappeared.

“So this is your dramatic return? Platform four, bad coffee, and you still picked the seat farthest from me.”

Through this character’s makeup. Theo’s elastic breath, associative wit and wish to make company easy turn a reunion into teasing. His sensitivity to rejection makes Ada’s chosen seat carry more emotional weight than its practical meaning alone.

The seven dramatic obligations

Time & Place. Late-night platform noise makes intimacy both private and exposed. The long bench physically displays the distance he cannot admit is painful.

Relationship. Ada is older, admired, and inconsistently available. Theo feels genuine affection and resentment, wanting her to initiate the apology he fears he cannot demand.

Emotional. Anticipation, anger, longing, and amusement coexist. The outward tone is teasing; the inner stakes are whether she has returned to stay.

Character. His bright energy and associative wit activate automatically. Watch how much faster he makes connections when waiting for the sister's response.

Historic. Contemporary travel and messaging habits make disappearance conspicuous: she could have written. The period creates an expectation of accessible contact without guaranteeing emotional availability.

Thematic. How many jokes can a person tell before they are allowed to say they were hurt?

Subtextual. 'You chose the seat away from me' means 'Are you leaving again?' The insult is a tiny request to be approached.

Proposed extension obligations

Playable objective. Invite Ada to cross the emotional distance without making a naked demand. Stakes. A short reunion may confirm the old pattern of unreliable closeness.

Obstacles. His fear of being needy; Her uncertain willingness to explain; Reflexive comic timing.

Social / status. Siblings are allowed familiar mockery, but shared humor is not the same as repaired trust. Ada decides whether to return; Theo tries to seize control through tone.

Knowledge / regulation. Theo does not know why Ada left or whether she intends to remain. Convert hurt into amusement, but leave audible space where the hurt can surface.

Current physiology. Eager forward lean, quick exhalations, an involuntary breath stop after the apparent punch line.

LAION situation dimensions for later codebook matching: SOCT · ENVI · TRGT · FATG

Partner-specific obligations

Older Sister / ada. Fondness mixed with perceived abandonment; his teasing tests whether their familiar rhythm still exists. Desired response: Ada acknowledges both the joke and the injury beneath it.

Ordered Acting Beats

BEAT 01 · tease

“So this is your dramatic return?”

Outer Acting Beat. Bright rhetorical lift; a fraction too energetic for a quiet platform.

Displayed affect. amusement.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Ada appears without explanation. Relief immediately shares space with accusation.

Attention / thought. Ada’s return without explanation. Unspoken fragment: “Laugh with me before you see what this costs.”

Sensation. Quick, eager breath. Memory. The long interval of scarce contact. Association. A reunion can be staged as a dramatic entrance.

Emotion. longing. Impulse and regulation. Perform surprise before she sees fear. Competing impulse. Ask directly why she disappeared.

Carryover. Begin from the character’s scene obligations; no earlier beat in this scene is supplied.

Vocal targets. Pitch Range (RANG) = 5; Playful Style (S_PLAY) = 5; Tempo (TEMP) = 4. Inherited whole-beat emotion labels (unassigned): Amusement, Longing.

BEAT 02 · charm

“Platform four, bad coffee,”

Outer Acting Beat. Rhythmic list, two light accents, breath timed like a setup.

Displayed affect. playfulness.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. The setting offers props for a comic escape route. His attention searches for shared safe material.

Attention / thought. The platform and bad coffee as shared comic material. Unspoken fragment: “We can still do this bit together.”

Association. An ordinary setting provides a route back to a familiar rhythm.

Emotion. guarded hope. Impulse and regulation. Make her laugh, then continue. Competing impulse. Abandon the joke and demand reassurance.

Carryover. Carry forward the earlier longing; the new stimulus changes its expression.

Vocal targets. Emphasis (EMPH) = 5; Smoothness (SMTH) = 4. Inherited whole-beat emotion labels (unassigned): Amusement.

BEAT 03 · accuse indirectly

“and you still picked the seat”

Outer Acting Beat. Tempo dips, still gains stress, smile-voice narrows.

Displayed affect. teasing with a sharp edge.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. He notices the physical gap again. The joke suddenly stops feeling harmless.

Attention / thought. The empty physical distance between their seats. Unspoken fragment: “You are still keeping away.”

Association. The chosen seat begins to resemble the previous absence.

Emotion. hurt. Impulse and regulation. Name her distance without pleading. Competing impulse. Plead rather than hide the request in teasing.

Carryover. Carry forward the earlier guarded hope; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 2; Tension (TENS) = 4; Vulnerability (VULN) = 3. Inherited whole-beat emotion labels (unassigned): Longing, Sadness.

BEAT 04 · invite

“farthest from me.”

Outer Acting Beat. Fall rather than a punch-line rise; silence after me is left unfilled.

Displayed affect. fragile joke.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Ada looks toward the empty space beside her. He hopes for movement and fears refusal.

Attention / thought. Ada’s glance toward the empty space. Unspoken fragment: “Come closer.”

Sensation. The breath waits with the unfinished possibility. Association. A small movement could answer the larger question.

Emotion. exposed affection. Impulse and regulation. Wait, finally. Competing impulse. Fill the silence with another joke before refusal becomes possible.

Carryover. Carry forward the earlier hurt; the new stimulus changes its expression.

Vocal targets. Vulnerability (VULN) = 5; Pitch Range (RANG) = 2. Inherited whole-beat emotion labels (unassigned): Longing.

Why this scene matters for controllable speech generation. The same comic reflex acquires different dramatic meanings: protective humor in one scene, chosen stillness in the other.

Scene 2 / The silence after the applause

EARLY EVENING, NEIGHBORHOOD ARTS HALL FOLLOWING A FUNDRAISER FOR A RETIRING MENTOR

Theo is asked to offer a toast to Sal, the mentor who gave him room to fail without withdrawing affection. For once, he chooses to stop being funny at precisely the moment the audience expects another joke.

“I used to fill every silence in this room. Tonight I can leave one here. Thank you for staying.”

Through this character’s makeup. Theo still has comic timing and social inventiveness, but Sal’s dependable presence permits a different use of them. A chosen pause becomes possible because belonging no longer has to be bought with a laugh.

The seven dramatic obligations

Time & Place. A full room follows an applause swell; the silence afterward feels almost physical. Unlike the railway platform, the audience is not departing, and Theo has time.

Relationship. Sal is mentor and reliable witness, not a replacement parent. Sal has modeled the ordinary act of staying through awkwardness.

Emotional. Gratitude, sadness at an ending, and a manageable fear of intimacy overlap. Theo shows warmth without hiding every tremor under comic energy.

Character. His inventiveness remains, but he directs it toward a simple reversal: turning a habit of filling silence into a gift of silence.

Historic. A contemporary public tribute normally rewards fluent gratitude and buoyancy; Theo gently violates the expectation of nonstop entertainment.

Thematic. Belonging can survive an interval in which nobody is performing.

Subtextual. 'Thank you for staying' means 'You proved I could be loved when I had nothing entertaining to offer.'

Proposed extension obligations

Playable objective. Offer Sal sincere acknowledgment without converting their relationship into a staged punch line. Stakes. The tribute may become one more act of concealment unless Theo takes a personal risk.

Obstacles. Audience expectation; Reflexive humor; Fear of losing the mentor’s stabilizing presence.

Social / status. Public speeches are a social performance, but a measured private truth can briefly reorganize the room. Theo has the microphone; Sal holds the deeper emotional authority.

Knowledge / regulation. He cannot control whether the audience experiences sincerity as uncomfortable. Allow vulnerability while retaining enough steadiness to complete the toast.

Current physiology. Breathing slows with a deliberate pause; a tremor appears on breath renewal, not continuously.

LAION situation dimensions for later codebook matching: SOCT · TRGT · ENVI · POSE

Partner-specific obligations

Retiring Mentor / sal. Warmth without debt-collection: Sal offered a model of durable presence. Desired response: Sal feels seen without being made responsible for Theo’s tears.

Ordered Acting Beats

BEAT 01 · acknowledge pattern

“I used to fill every silence in this room.”

Outer Acting Beat. Starts with familiar lift, then deliberately releases the final word.

Displayed affect. wry openness.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Applause fades and the room expects his familiar rhythm. He notices an opportunity to do something different.

Attention / thought. Applause fades; the room waits for his familiar comic response. Unspoken fragment: “I do not have to rescue this moment.”

Memory. His old habit of filling the room’s silences. Association. The expected joke can become an opportunity to choose differently.

Emotion. gentle embarrassment. Impulse and regulation. Make a joke, then refrain. Competing impulse. Provide the next laugh automatically.

Carryover. Begin from the character’s scene obligations; no earlier beat in this scene is supplied.

Vocal targets. Playful Style (S_PLAY) = 2; Tempo (TEMP) = 3; Vulnerability (VULN) = 4. Inherited whole-beat emotion labels (unassigned): Contemplation, Amusement.

BEAT 02 · offer space

“Tonight I can leave one here.”

Outer Acting Beat. Slow, narrow melodic movement; visible imagined pause after here.

Displayed affect. stillness.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. No one interrupts the pause. The silence becomes safety rather than abandonment.

Attention / thought. Nobody interrupts the pause. Unspoken fragment: “They are still here.”

Sensation. Breathing slows. Association. Silence can mean continued company rather than abandonment.

Emotion. gratitude. Impulse and regulation. Trust that the room will remain. Competing impulse. Explain or decorate the silence before it becomes uncomfortable.

Carryover. Carry forward the earlier gentle embarrassment; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 2; Pitch Range (RANG) = 2; Warmth (WARM) = 5. Inherited whole-beat emotion labels (unassigned): Contentment.

BEAT 03 · thank

“Thank you”

Outer Acting Beat. Soft onset; vowels free from the usual comic smile.

Displayed affect. tenderness.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Sal meets his gaze. Feeling gathers without needing to become spectacle.

Attention / thought. Sal’s returned gaze. Unspoken fragment: “Say it to Sal.”

Memory. Sal offered room to fail without withdrawing affection.

Emotion. love and gratitude. Impulse and regulation. Speak directly. Competing impulse. Turn gratitude into a performance for the whole room.

Carryover. Carry forward the earlier gratitude; the new stimulus changes its expression.

Vocal targets. Vulnerability (VULN) = 5; Warmth (WARM) = 5. Inherited whole-beat emotion labels (unassigned): Contentment, Longing.

BEAT 04 · honor

“for staying.”

Outer Acting Beat. A fine break at staying is allowed, then no commentary.

Displayed affect. acceptance.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. The last word names what mattered all along. Relief and grief briefly coexist.

Attention / thought. The meaning of staying as the final word arrives. Unspoken fragment: “That was what mattered.”

Sensation. A small instability on breath renewal may be allowed. Memory. The mentor’s durable presence through awkwardness. Association. Relief at belonging and sadness at an ending coexist.

Emotion. sad gratitude. Impulse and regulation. Finish without a rescue joke. Competing impulse. Add a rescue joke after the admission.

Carryover. Carry forward the earlier love and gratitude; the new stimulus changes its expression.

Vocal targets. Tension (TENS) = 2; Disfluency (DFLU) = 2; Vulnerability (VULN) = 5. Inherited whole-beat emotion labels (unassigned): Sadness, Contentment.

Why this scene matters for controllable speech generation. The same comic reflex acquires different dramatic meanings: protective humor in one scene, chosen stillness in the other.

Evelyn Hart

Authored character design · two scenes · eight ordered beats. Every beat contains one Experience Chunk. Original portraits and scene detail are retained; new inner-fragment additions are fictional conditioning.

She protects truth by narrowing her voice, until truth requires it to open.

MUNICIPAL RECORDS CUSTODIAN • FICTIONAL ENGLISH-SPEAKING INDUSTRIAL CITY IN 1934

Long-range identity: She protects truth by narrowing her voice, until truth requires it to open.

Character Makeup / four-domain portrait

Physical disposition

Evelyn uses physical restraint as both occupational practice and social armor. She stands as if a record drawer might slide open unless she keeps the entire room in balance. Fine-motor tasks have trained her to favor small exact gestures; after years in poorly heated archives, she takes an extra moment to warm her hands before handling delicate paper. Her speaking voice is grounded and economical, neither ceremonially grand nor theatrically frail. When unsettled, her breath remains disciplined but the release of sentence-final consonants becomes unusually careful. An unusual condition of her performance is that the audience can hear the amount of effort she spends not appearing distressed. Her physical world is constructed for this fiction; no audible trait diagnoses arthritis or other disease.

Voice consequence. Measured air, lower-mid habitual placement, quiet friction before difficult names, and sentences that land with exact boundaries.

Countertendency. Strongly maintained stillness makes a small vocal fracture narratively large.

Emotional disposition

Evelyn's emotional point of view is watchful dignity. She is capable of affection and moral outrage but fears both will be used to discount her competence in public. At work she converts outrage into impeccable citation; in private she can be unexpectedly gentle with someone who has been excluded or publicly dismissed. Her tenderness is not demonstrative, and she interprets direct demands for reassurance as a threat to the other's autonomy. The cost is loneliness disguised as principled independence. When evidence implicates someone she loved, grief and loyalty do not disappear; they form a contradictory emotional obligation that her procedural language can contain only for a short time. Her most powerful shift is from formal certainty to admitted uncertainty.

Voice consequence. Controlled gravity, little ornament, warmth expressed by permitting unfinished thought rather than by sudden sweetness.

Countertendency. She tends to substitute precision for emotional admission precisely when her values are most challenged.

Intellectual disposition

Evelyn is a meticulous source critic. She notices chronological gaps, provenance errors, and the difference between an assertion and something a surviving document can actually support. Her habitual syntax makes explicit the limits of knowledge: she prefers 'the record suggests' to 'everyone knows.' She is also cognitively proud, sometimes mistaking caution for neutrality when it is partly fear of personal exposure. She can hold two incompatible interpretations in mind without rushing to dissolve them, yet she struggles to bring that tolerance of ambiguity into family history. Under public attack, her reasoning becomes even clearer, while under private shock it fragments into a list of facts that no longer answers the human question in front of her.

Voice consequence. Distinctive stress on evidential words, patient clause boundaries, and suddenly shortened syntax when the world outruns analysis.

Countertendency. The ability to distinguish claims from evidence can become a defensive refusal to state her own stake.

Psychological disposition

Evelyn has built her identity around being difficult to intimidate and impossible to buy. She regards the archive as a refuge from the social rituals that rank people by family ties, reputation, and access. Yet her own reputation matters to her, particularly because she remembers how often she had to defend the legitimacy of her work. She has a powerful moral self-image: keeper of the record rather than participant in the story. The suppressed contradiction is that preserving evidence can expose people she loves, including herself. In public her protective self demands immaculate calm; in a trusted one-to-one exchange a less practiced part wants to share both the evidence and the cost of carrying it. That second voice is present even before she permits it audible form.

Voice consequence. Public certainty may be a shield; private confession gains force through reduced structure and unmistakable hesitations.

Countertendency. She can become unfair to close partners when she mistakes their empathy for an invitation to compromise standards.

Scene 1 / Before the civic committee

1934, MUNICIPAL HEARING ROOM, AFTERNOON; COLLEAGUES AND PRESS LISTENING

Chair Bell pressures Evelyn to certify that missing pages prove misconduct by an absent clerk. Evelyn is unsure, dislikes the clerk's politics, and yet refuses to substitute her preference for documentation. A junior assistant, Miles, watches her risk her position.

“The ledger contains a gap. I will not fill it with a convenient answer. Ask me what the gap proves.”

Through this character’s makeup. Evelyn’s physical restraint, evidential syntax and moral self-image make precise refusal her form of courage. Anger and fear remain active while public dignity constrains how she can express them.

The seven dramatic obligations

Time & Place. The municipal hearing room magnifies restrained sounds; the distance to the chair requires projection without theatrical amplitude. An open ledger sits between her and the committee.

Relationship. Bell possesses institutional authority but not Evelyn's technical expertise. Miles is a junior colleague for whom Evelyn wants to model intellectual courage without romanticizing its cost.

Emotional. She feels anger at coercion and fear of dismissal, while showing controlled professional conviction. The slight tremor she resists matters more than any open accusation.

Character. Her meticulous intellectual makeup becomes action: speak only what the pages justify. Her psychological need to remain incorruptible gives precision a moral charge.

Historic. The fictional 1934 social environment places weight on hierarchy, public decorum, paper evidence, institutional reputation, and formal address. These are authored period circumstances, not a history textbook claim.

Thematic. An archive is not honest because its keeper sounds certain; it is honest when uncertainty cannot be bullied into a convenient story.

Subtextual. 'Ask me what the gap proves' means 'I will not become your instrument, even if you can punish me.'

Proposed extension obligations

Playable objective. Prevent an unsupported conclusion from entering the official record. Stakes. Both the clerk’s reputation and the credibility of the archive are at risk.

Obstacles. Committee pressure; Risk to employment; Her own dislike of the absent clerk; Her need for public dignity.

Social / status. Deference is expected in the hearing; dissent must travel through controlled formal language. Bell controls proceedings; Evelyn controls documentary expertise.

Knowledge / regulation. She knows pages are missing, not why they are missing. Constrain anger into factual precision, without flattening the ethical urgency.

Current physiology. Rigid posture at the lectern, economical breath, controlled jaw effort under public attention.

LAION situation dimensions for later codebook matching: SOCT · TRGT · ENVI · POSE

Partner-specific obligations

Committee Chair / chair_bell. Adversarial respect: she follows formal address but refuses epistemic submission. Desired response: Bell stops presenting conjecture as established fact.

Junior Archive Assistant / miles. Professional mentorship; she wants him to learn that accuracy sometimes has a price. Desired response: Miles sees courage as method rather than bravado.

Ordered Acting Beats

BEAT 01 · distinguish

“The ledger contains a gap.”

Outer Acting Beat. Even delivery with strong landing on gap; low dynamic range.

Displayed affect. precision.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Bell calls the missing pages obvious proof. Evelyn separates the fact from the imposed interpretation.

Attention / thought. Bell’s claim that missing pages are obvious proof. Unspoken fragment: “Name only what the ledger contains.”

Sensation. Controlled jaw effort under public attention. Association. A missing record and a convenient explanation are different objects.

Emotion. contained anger. Impulse and regulation. Correct without inflaming. Competing impulse. Let dislike of the clerk substitute for documentary evidence.

Carryover. Begin from the character’s scene obligations; no earlier beat in this scene is supplied.

Vocal targets. Articulation Clarity (CLRT) = 5; Tempo (TEMP) = 2; Emphasis (EMPH) = 4. Inherited whole-beat emotion labels (unassigned): Concentration.

BEAT 02 · refuse

“I will not fill it”

Outer Acting Beat. Shorter breath before will; firm but not loud onset.

Displayed affect. resolve.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Bell invites a shortcut that would end her discomfort. Personal risk becomes immediate.

Attention / thought. The offered shortcut and its immediate personal cost. Unspoken fragment: “I will not do that.”

Sensation. A shorter preparatory breath. Association. Agreement could end the uncomfortable encounter.

Emotion. fear. Impulse and regulation. Refuse clearly. Competing impulse. Protect her position by accepting Bell’s interpretation.

Carryover. Carry forward the earlier contained anger; the new stimulus changes its expression.

Vocal targets. Tension (TENS) = 4; Vulnerability (VULN) = 2; Emphasis (EMPH) = 5. Inherited whole-beat emotion labels (unassigned): Fear, Concentration.

BEAT 03 · expose

“with a convenient answer.”

Outer Acting Beat. Convenient receives controlled acid; ending avoids a theatrical sneer.

Displayed affect. severe restraint.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. The committee shifts impatiently. Resentment flickers under technical criticism.

Attention / thought. The committee’s impatience. Unspoken fragment: “Let them hear what this shortcut does.”

Association. The proposed answer is convenient because it serves somebody.

Emotion. anger. Impulse and regulation. Let Bell hear the ethical indictment without naming him. Competing impulse. Turn controlled criticism into an overt personal attack.

Carryover. Carry forward the earlier fear; the new stimulus changes its expression.

Vocal targets. Tempo (TEMP) = 2; Pitch Range (RANG) = 2; Emphasis (EMPH) = 5. Inherited whole-beat emotion labels (unassigned): Anger.

BEAT 04 · challenge

“Ask me what the gap proves.”

Outer Acting Beat. Slight forward projection and clear question-shape, space after proves.

Displayed affect. invitation with authority.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. She chooses to make the limit of evidence explicit. Fear reorganizes as purposeful resolve.

Attention / thought. A question that can distinguish a gap from proof. Unspoken fragment: “Ask what I can actually answer.”

Association. Naming the limit of evidence can itself be an action.

Emotion. courage despite fear. Impulse and regulation. Invite a question she can answer truthfully. Competing impulse. Retreat after the refusal to avoid further exposure.

Carryover. Carry forward the earlier anger; the new stimulus changes its expression.

Vocal targets. Articulation Clarity (CLRT) = 5; Stance (STNC) = 5; Vulnerability (VULN) = 3. Inherited whole-beat emotion labels (unassigned): Concentration.

Why this scene matters for controllable speech generation. Formal precision remains part of the voice identity, but its function changes from public resistance to private confession.

Scene 2 / The final page

1934, UNHEATED ARCHIVE STORAGE ROOM, EVENING; PAPERS SPREAD ON A SMALL TABLE

A private letter suggests that Esther's father knowingly altered a municipal record. Evelyn recognized his writing immediately, concealed that recognition for an hour, and now must tell Esther before the finding becomes public.

“This is your father's handwriting. I recognized it at once. I pretended I did not. Read the final page.”

Through this character’s makeup. The archivist’s disciplined voice can identify the handwriting, yet precision cannot absolve her delay. Loyalty, shame and her identity as keeper of the record make the confession harder than the factual disclosure.

The seven dramatic obligations

Time & Place. The enclosed archive reduces projection to intimate speech. Cold air, paper dust, and a close listener turn normally impersonal evidence into an immediate bodily presence.

Relationship. Esther is a brilliant younger colleague who trusts Evelyn's honesty. Evelyn fears that confessing her own delay will feel like betrayal rather than care.

Emotional. Protectiveness, guilt, dread, and grief compete. She tries to remain factual but can no longer separate evidence from the listener's impending loss.

Character. The disciplined archivist becomes an unreliable narrator of her own choices. Intellectual precision survives, but her psychological self-image is punctured by the admission of concealment.

Historic. Reputation and kinship in this fictional 1934 setting create professional and family consequences. The paper letter is physically present and cannot be dismissed as a rumor.

Thematic. The cost of telling the truth is greatest when the person harmed is not an abstraction.

Subtextual. 'I pretended I did not' means 'I failed you for one hour because I could not bear to be the messenger.'

Proposed extension obligations

Playable objective. Tell Esther the material fact and own the delay without making her carry Evelyn’s guilt. Stakes. The assistant may lose trust in both her family story and her mentor.

Obstacles. Esther’s likely distress; Her loyalty to Esther; Her shame at postponing disclosure.

Social / status. Privacy permits direct speech but does not remove the ethical duty of disclosure. Evelyn holds information and workplace authority; Esther owns the emotional consequences.

Knowledge / regulation. Handwriting appears recognizable; the full meaning of the letter still requires interpretation. Stop using procedural formality to escape responsibility; do not force emotional resolution.

Current physiology. Voice moved closer and lowered; breath suspension before her admission.

LAION situation dimensions for later codebook matching: SOCT · TRGT · ENVI · CLMT

Partner-specific obligations

Junior Colleague / esther. Mutual regard with an asymmetry of experience, authority, and access to evidence. Desired response: Esther receives the truth and can decide what it means to her.

Ordered Acting Beats

BEAT 01 · identify

“This is your father's handwriting.”

Outer Acting Beat. Lower register, unhurried father’s, slight breath reserve.

Displayed affect. quiet sobriety.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Esther asks what makes the letter significant. Professional fact becomes intimate harm.

Attention / thought. Esther’s question about the letter. Unspoken fragment: “Begin with the fact.”

Sensation. Cold air and a close listener in the archive. Memory. Recognition of her father’s handwriting. Association. An impersonal document becomes a family consequence.

Emotion. dread and care. Impulse and regulation. Place one verifiable fact on the table. Competing impulse. Delay the disclosure to protect Esther from the immediate shock.

Carryover. Begin from the character’s scene obligations; no earlier beat in this scene is supplied.

Vocal targets. Tempo (TEMP) = 2; Warmth (WARM) = 3; Tension (TENS) = 4. Inherited whole-beat emotion labels (unassigned): Sadness, Concentration.

BEAT 02 · admit

“I recognized it at once.”

Outer Acting Beat. Exact rhythm, little inflection; the final word is not defended.

Displayed affect. plain candor.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Esther notices that Evelyn is not surprised. The delay becomes impossible to conceal.

Attention / thought. Esther notices that Evelyn is not surprised. Unspoken fragment: “Do not qualify when I knew.”

Memory. The recognition happened at once, an hour earlier.

Emotion. shame. Impulse and regulation. Name prior knowledge without qualification. Competing impulse. Explain the delay before admitting prior knowledge.

Carryover. Carry forward the earlier dread and care; the new stimulus changes its expression.

Vocal targets. Articulation Clarity (CLRT) = 4; Pitch Range (RANG) = 1; Vulnerability (VULN) = 4. Inherited whole-beat emotion labels (unassigned): Sadness.

BEAT 03 · confess

“I pretended I did not.”

Outer Acting Beat. Tiny silence before pretended; breath catches and then resumes normally.

Displayed affect. vulnerability.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Evelyn confronts her own breach of trust. Self-concept and observed action collide.

Attention / thought. Her own choice to conceal recognition. Unspoken fragment: “I did conceal it.”

Sensation. A brief suspension of breath. Memory. The hour of pretending not to recognise the handwriting. Association. Her claim to honesty collides with what she did.

Emotion. guilt. Impulse and regulation. Confess instead of hiding behind evidence. Competing impulse. Hide behind procedure or uncertainty in the document.

Carryover. Carry forward the earlier shame; the new stimulus changes its expression.

Vocal targets. Disfluency (DFLU) = 3; Respiration (RESP) = 4; Vulnerability (VULN) = 5. Inherited whole-beat emotion labels (unassigned): Sadness, Fear.

BEAT 04 · yield

“Read the final page.”

Outer Acting Beat. No command snap despite imperative grammar; warm, low, room left afterward.

Displayed affect. restrained invitation.

One Experience Chunk · authored inner stream

Stimulus → experienced effect. Esther reaches toward the document. Evelyn lets agency return to Esther.

Attention / thought. Esther reaches toward the page. Unspoken fragment: “Let her read and decide.”

Association. Returning the evidence can return agency to its reader.

Emotion. sorrow with respect. Impulse and regulation. Offer evidence; do not decide her response. Competing impulse. Manage Esther’s interpretation or ask her to relieve Evelyn’s guilt.

Carryover. Carry forward the earlier guilt; the new stimulus changes its expression.

Vocal targets. Tension (TENS) = 2; Warmth (WARM) = 5; Tempo (TEMP) = 2. Inherited whole-beat emotion labels (unassigned): Sadness.

Why this scene matters for controllable speech generation. Formal precision remains part of the voice identity, but its function changes from public resistance to private confession.

Build, generate and review

Use the Script Analysis System to prepare a virtual actor’s inputs. Use the Performance Annotation System for the resulting performance or a permitted reference recording; keep intended conditions and supported observations in separate records.

  1. Read and locate the source. Preserve text, roles, scene boundaries and source locators. Mark direct source facts, interpretation and new inventions.
  2. Build the characters. Write all four makeup portraits, contradictions and possible realization consequences. Add only permitted reference media, with a clear intended use.
  3. Interpret the scene. Fill all seven Dramatic Obligations for each character, including Emotional. Connect each scene’s meaning to that character’s makeup and partner relations.
  4. Order the beats. Anchor dialogue exactly where it exists. State concrete actions or silence where there is no speech. Add usable timing and delivery cues.
  5. Write one chunk per beat. Specify the fictional inner stream and its continuity. Keep thoughts unspoken unless the source actually gives them as dialogue. Remove inner detail that does not help the performance.
  6. Validate and generate. Check the schema, IDs, text anchors, ordering and availability of assets. Then send the selected character, scene and beat package to a generator that supports the desired modality.
  7. Inspect the rendered result. Preserve the actual transcript and timing. Use the Performance Annotation System to describe available evidence and task fit. Compare variants without treating the authored inner stream as an observed psychological fact.
  8. Revise the design deliberately. If the result is too broad, change a specific cause, action, timing cue or local realization. Preserve other conditions when testing a controlled alternative.

Checks that matter before a model call

  • Every scene has all seven obligations and a character-specific interpretation.
  • Every beat has exactly one Experience Chunk object, not zero, not a list of independent chunks.
  • Every non-null text anchor occurs in the scene’s source text. Repeated matching words need an occurrence or span distinction in production data.
  • Beat IDs are unique, orders are consecutive and carryover points to the preceding beat in the same scene.
  • A non-speaking beat has a concrete action or silence; missing timing is null, not a fabricated timestamp.
  • Speaker or avatar identity references, permitted media, modality requests and codebook choices are explicit.
  • Prompted emotions, desired vocal targets, actual output and observer judgements remain separate.

Test whether the hierarchy helps

Compare a simple emotion prompt with full character-and-scene conditioning, then with the same design plus local beats and chunks. Hold the source text and output identity stable. Ask whether the response remains character-consistent, locally responsive, intelligible, restrained where needed and faithful to the intended action. Test the inner chunk as a design variable; it is not ground truth about any model’s mental state.

Research foundations

The generation design draws on acting concepts while extending their representation for virtual actors. The distinctions below preserve the original research attribution and its limits.

What Eric Morris actually contributes - and where the proposed annotation architecture begins.

Character Obligation, not cosmetic makeup

In Schuhmann’s structured knowledge-base entry, the Character Obligation concerns building a role across four broad domains: physical, emotional, intellectual, and psychological. The actor compares themselves with the role, uses imaginative backstory where useful, and makes personally usable choices that stimulate the role’s behavior. This is the closest directly verified framework for the “character makeup” requested here. The retrieved material does not establish a canonical, numbered Morris checklist of subfields under those four domains; the detailed subfields in this report are an original operationalization. [02]

The same knowledge base describes characterization as a process of bringing the role into the actor’s available experience rather than impersonating a separate person. Its four evidence channels are what a writer says, what others say about the character, what the character says about themselves, and what they do. It particularly points to subpersonalities, external stimuli and affective-memory work as rehearsal routes. That matters conceptually, but a retrospective audio label must never assert which private rehearsal route the human performer used. [03, 05-07]

Three research claims that should remain distinct

ORIGIN JUSTIFIED CLAIM WHAT THIS DOCUMENT ADDS
Morris / knowledge base Four role domains; seven obligations; choice and choice approach; being; moment-to-moment responsiveness Detailed portrait grammar, explicit annotation uncertainties, scene and beat object model
Acting practice Given circumstances, objectives, obstacles, playable actions, behavior and backstory [19, 20, 22] Partner-state arrays, social contracts, epistemic gaps, masking, long-form actor questions
LAION vocal taxonomies VoiceNet speech scales, EmoNet emotion candidates, VocalBurst classes and Situation dimensions [15-18, 23] Controlled tags placed below the dramatic interpretation rather than confused with it

Morris: from character condition to lived response

The repository’s description of the Morris craft explicitly separates Obligations (realities the actor must create), Choices (personally meaningful stimuli selected to help make them real), and Choice Approaches (methods for working with those choices). In the present specification, the output should describe the dramatic construct and its perceptible manifestation; the private choice or technique can be included only when supplied by the actor or director. [04]

Earlier terminology. The original research PDF used “Moment-to-Moment Acting State” for a proposed annotation layer. This generation specification expresses the third layer as ordered Acting Beats with one nested Experience Chunk each; the earlier term does not name a fourth layer. Morris’s moment-to-moment principle is not a fixed list of emotional beats: it emphasizes unpremeditated responsiveness to live stimuli. A generated control sheet can plan conditions for responsiveness , but should not claim to reproduce the performer’s genuine lived experience. [08-10]

Responsiveness, regulation and the local phrase

The Morris-derived responsive sequence is stimulus → effect → response → expression. In the generation design, this motivates a causal connection between the outer beat and its authored inner chunk. It does not mean that a fixed control sheet reproduces a human performer’s unpremeditated lived experience. A local turn may arise from a partner’s reaction, a new thought, an involuntary body response or a decision to mask an emotion. [09, 10]

The original phrase questions remain useful for authoring: What just happened? What did the figure attend to? What could it mean to them? What urge appeared? What did they decide to do to the listener? Which contradictory urge did they restrain? What changed in breath, rhythm, melody, articulation, loudness and silence? What would another plausible dramatic interpretation be?

Felt state, displayed affect and vocal realization remain separate. James Gross’s work on expressive suppression versus reappraisal motivates the distinction between inner experience, expressive behavior and physiology. This is conceptual motivation for the fictional design; it is not permission to diagnose an inner emotion from a clip or a claim that a generator experiences that emotion. [21]

Vocal vocabulary and testing

A crosswalk to audio vocabularies: complementary measurements, never a substitute for causal dramaturgy.

The inherited sources cite a 57-dimension base VoiceNet rubric and a separate 59-dimension voice-taxonomies version, 40 EmoNet labels, 82 VocalBurst categories and 289 Situation scenarios in 11 dimensions. These are source-specific inventories. In Humaneness PerformanceNet, the PerformanceNet Taxonomy has 82 descriptive fields of its own; its count is unrelated to that older 82-category VocalBurst inventory. Use the pinned VoiceNet and EmoNet codebooks and the shared Situation Taxonomy volume for the applicable current representation. Preserve historical references rather than silently equating differing inventories. [15–17]

SCENE / BEAT FIELD EXTERNAL TAXONOMY EXAMPLE ASSOCIATION
vocal_realization VoiceNet, version pinned TEMP, CLRT, TENS, RESP, RANG, VULN, WARM, EMPH
displayed_affect EmoNet (expressed, not proven felt) Fear, Sadness, Concentration, Longing, Amusement
non-speech events VocalBurst sigh, breath, gasp, laughter (use exact leaf IDs when available)
place / body / norms Situation Taxonomy POSE, ACTV, TRGT, SOCT, ENVI, HLTH, GEAR, CLMT, SUBST, FATG, PAIN
difficult blend Acting Edge Cases masked emotion, contradictory signals, physical distress

Example coding note: The fictional records use valid LAION dimension codes and illustrative desired VoiceNet 0-6 levels, not measured predictions. EmoNet entries are candidate emotion names for a director to verify against the chosen codebook. Situation references are dimension-level tags, not verified identifiers for exact taxonomy leaf entries. [15-17]

Generation tests and optional reference analysis

The Script Analysis package is assembled forward from source material and authored design. Optional analysis of permitted reference performances can inform voice or movement choices, but a recording does not supply verified private memories, motives or inner experience. Generate controlled variants from the same source text, then use the Annotation System to encode the rendered evidence separately. The following tests retain the original research questions as proposed tests, not demonstrated results.

CONTROL PROBLEM PROPOSED TEST
Identity preservation Keep speaker voice identity stable while scene needs and subtext change.
Scene transfer Same character, new obligation: controlled leader in crisis → accountable mother afterward.
Disentanglement Same sentence, held text fixed: fear masked as reassurance vs directly displayed fear.
Counterfactual consistency Same situation, different Character Makeup: comedic deflection vs procedural containment.
Time-local precision Anchor an apology or impulse change to a word span without overacting the whole line.
Calibration Check that weakly supported private motives are not elevated to ground truth.
Human preference Experienced listeners rate interpretive coherence, naturalness, intelligibility and restraint.

Sources and implementation files

Naming continuity. Humaneness Script Analysis System is the current name of the framework previously presented as Actor-State Atlas. The four supplied illustrations and inherited bibliographic titles retain that historical name. Their content is integrated into the same three-layer architecture; they do not identify a separate current system. Existing schema filenames, field names, source codes and stable HTML IDs remain unchanged for traceability.

What is supplied with this edition

The new generation schema and example package are in the editable data/ folder. The examples preserve the original three fictional character dossiers and reorganize all twenty-four beats around one outer specification and one Experience Chunk. The revision audit records retained source units, relocated research material and newly authored inner fields.

Legacy companion availability. The original PDF named actor_state_dictionary_v1.schema.json and actor_state_illustrative_examples_v1.json, but those files were not supplied. Its printed hierarchy sketch used explanatory ellipses. This edition replaces that sketch in the main reading path with valid JSON and supplies its own separately named schema and examples; it does not reconstruct or claim validation of the missing legacy files.

Illustrated source and editorial additions

Actor-State Atlas: Three Layers Illustrated, supplied three-page research proposal, 17 October 2026, provides the four reproduced illustrations and the explicit “one nested Experience Chunk per beat” architecture. The source research standard is dated 15 October 2026. This generation revision incorporates the requested source-text → character → scene → beat-and-chunk workflow, complete machine companions and new authored inner fragments while preserving the original portraits, scene obligations and vocal directions.

The images are reproduced as supplied. Their references to inner experience describe fictional character conditioning. Their multimodal outputs are proposed applications. No generated recording, empirical controllability result or AI-consciousness claim is supplied by this document.

Source fidelity and design boundaries

The authorial interpretation of Morris is based on a secondary, machine-extracted knowledge base that cites seven books. Some entries contain unknown page references, so the best way to settle fine distinctions in a publication-quality scholarly edition would be to consult the printed works directly. The public acting resources are used for context, not to collapse all approaches into one acting tradition.

There is no claim in this document that Eric Morris, LAION, Schuhmann, or any acting school endorses the proposed dictionary or sample interpretations. The sample personae and their dialogue were invented for this project; no sourced dramatic dialogue was copied.

Inherited references

The heading and citations below are inherited from the source. Link destinations have been recovered from embedded PDF hyperlinks where printed URLs were abbreviated. This is a provenance list, not a fresh verification of the linked resources.

[01] Schuhmann, C. Eric Morris Acting System: knowledge-base overview.

https://github.com/christophschuhmann/eric_morris-method-acting-system

[02] Schuhmann, C. Character Obligation: four-part framework and rehearsal procedure.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/character-obligation.html

[03] Schuhmann, C. Characterization: four sources of evidence and recommended approaches.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/characterization.html

[04] Schuhmann, C. Obligations: seven obligations, choices, and choice approaches.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/obligations.html

[06] Schuhmann, C. Externals: outside-in characterization through objects, people and animals.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/externals.html

[07] Schuhmann, C. Imaging: multimodal imaginative process for choices.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/imaging.html

[09] Schuhmann, C. Moment-to-Moment: responsiveness rather than premeditation.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/moment-to-moment.html

[11] Schuhmann, C. Relationship Obligation: identity of partner and feeling toward partner.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/relationship-obligation.html

[12] Schuhmann, C. Historic Obligation: period, conventions and lived reality.

https://github.com/christophschuhmann/eric_morris-method-acting-system/blob/main/concepts/historic-obligation.html

[14] Morris, E. Official school / introductory interview with the creator.

https://ericmorris.com/

[15] LAION-AI. Voice Taxonomies: VoiceNet, EmoNet, VocalBurst, Acting Edge Cases.

https://github.com/LAION-AI/voice-taxonomies

[16] LAION-AI. Situation-Driven Acting: 11 dimensions and 289 situations.

https://projects.laion.ai/Voice-Acting-Pipeline/path_sit_situation.html

[17] LAION-AI. VoiceNet base taxonomy: 57 dimensions and ordinal rubrics.

https://projects.laion.ai/Voice-Acting-Pipeline/voicenet_taxonomy.html

[18] LAION-AI. Acting Challenge dataset: Eric Morris-inspired challenges.

https://projects.laion.ai/Voice-Acting-Pipeline/path_ac_acting_challenge.html

[19] Backstage. Character Development: actor-oriented backstory, relationship and vocalization.

https://www.backstage.com/magazine/article/character-development-for-actors-76083/

[20] ACT Broward. Uta Hagen: Six Steps for Building a Character.

https://www.actbroward.org/uta_hagen.php

[21] Gross, J. J. (1998). Antecedent- and response-focused emotion regulation. JPSP 74, 224-237.

https://pubmed.ncbi.nlm.nih.gov/9457784/

[22] Backstage. Guide to the Stanislavsky technique: given circumstances, action and objectives.

https://www.backstage.com/magazine/article/the-definitive-guide-to-the-stanislavsky-acting-technique-65716/

[23] LAION-AI. Acting Edge Cases and intended acoustic signatures.

https://github.com/LAION-AI/acting-edge-cases