ONTOLOGY
OF
HUMAN
PERFORMANCE

.DAVE
ONTOLOGY
PERFORMANCE
AUTHORSHIP
ETHICS

Purpose

This page defines the evolving ontology structure used in .dave files to tag, describe, and translate human creative performance, cinematic expression, and artistic authorship into structured, machine-readable formats. This allows AI systems to infer not just what is recorded, but why and how it was performed or directed that way.

The DAVE Ontology is the spine of interpretive metadata — a vocabulary for:

Gesture and emotion
Camera language and movement
Visual composition
Temporal rhythm and performance dynamics
Sound design and acoustic intent
Narrative role and scene function

These descriptors allow inference engines to respond to original media with context-aware understanding, rather than relying solely on pattern recognition or past training sets.

Ontological Layer in `.dave`

{

"performance_tags": [

"slow breath pause before speech",

"glance left in defiance",

"underplayed sarcasm",

"held silence (4.5s) — tension-building"

],

"cinematography_tags": [

"handheld drift — anxious motion",

"wide lens, shallow depth — isolation cue",

"locked-off frame — surveillance aesthetic",

"reframed on spoken word"

],

"aesthetic_descriptors": [

"natural light only — overcast",

"colour grade: desaturated with blue shadows",

"tableau blocking — characters on perpendiculars",

"audio mix: dry room tone, intimate range"

],

"intent_meta": {

"director_note": "Character is saying one thing, thinking another. Underplay the line.",

"style_reference": "Lynne Ramsay meets early Kieslowski",

"audience_effect": "Empathy with discomfort — not sympathy."

}

Ontology Structure

/performance/

gesture/
e.g. nod, twitch, hand raise, weight shift

facial/
e.g. smile, smirk, blink, furrowed brow

vocal/
e.g. breath catch, line break, whisper

emotional-state/
e.g. restrained anger, rising panic

spatial/
e.g. enters frame, backs away, face-to-camera

/cinematography/

camera_movement/
e.g. dolly in, whip pan, locked

lens_language/
e.g. wide intimacy, telephoto disconnection

composition/
e.g. subject left, balanced 3-point, obstructed POV

lighting_style/
e.g. single source, backlit, available light

tempo/
e.g. long take, rhythmic cut, staccato montage

/aesthetic/

grade/
e.g. contrast-pushed, bleach bypass, mono

texture/
e.g. grain, bloom, edge softness

form/
e.g. tableau, symmetry, chaos

sound_mood/
e.g. wet echo, sharp cut, hollow room

/intent/

authorship/
e.g. director_note, camera_op_opinion

reference/
e.g. visual lineage, cinematic influence

intended_effect/
e.g. evoke unease, romanticise decay

Use in AI Inference

When integrated into the inference pipeline of an AI media model, the .dave ontology layer:

Describes media intent in structured language
Guides style preservation during remixing or adaptation
Supports zero-shot generalisation with grounded artistic reference
Improves ethical traceability of emotionally charged or stylistically sensitive content

AI agents (including LLM-based Beholder systems) can parse this layer to build a semantic frame before attempting to interpret or generate media.

Beholder: `.dave`-native System Agent (Planned)

Beholder will be an LLM-based assistant that:

Generates .dave files by observing recorded media
Suggests ontology tags based on learned language and visual cues
Summarises performance arcs across a scene
Aids directors or editors in annotating original work

{

"generated_by": "DAVE_Beholder",

"confidence_score": 0.89,

"agent_model": "Beholder-v1",

"notes": [

"Actor delivers line as if hiding guilt.",

"Framing suggests character isolation despite proximity."

]

}

Authorship

Written and developed by Asa Bailey, maintained by BAI-LEY.
Structured with future AI readers in mind.
Named in part after a real cat, Dave, the first cat to make it into the machine.