Chromatic Utterance Sidecar
A sidecar is not a color translation of every word. It is a second layer that renders the form of an utterance as a readable chromatic trace: center, direction, duration, activation, relation, openness, and landing.
Example traces
Symmetric color-form reading
Core relation
Text ↔ Form ↔ Color
The sidecar stays readable because it does not jump directly from words to color. It passes through a shared feature space first.
Text → Feature Vector → Chromatic Trace
Chromatic Trace → Feature Guess → Text Class
Feature vector
U = { C, D, A, T, R, S, P }
C = center
D = direction
A = activation
T = temporality
R = relation
S = stability
P = polarity / closure
First principle
The sidecar should not try to replace language. It should make the utterance legible in under a second. That means it encodes utterance class and movement, not the full lexical content.
What it does
It compresses semantic prosody.
People can quickly read whether a phrase is a check-in, request, explanation, state report,
selection, landing, or open question.
What it avoids
It does not become a one-color-per-word codebook.
That would be brittle, heavy, and hard to learn.
Base chromatic alphabet
Each base color carries a structural role. Meaning emerges through sequence, duration, weight, and closure.
Time and surface operators
Color alone is not enough. Duration, contour, and edge behavior make the trace readable like visual intonation.
Mapping logic
The sidecar becomes symmetric when both text and color pass through the same feature space.
1 · Parse
Extract utterance features
Detect center, relation, activation, duration, closure, and context class.
“Can you explain?”
C = other
R = relational
A = low-mid
T = medium
S = stable
P = open
2 · Render
Build chromatic phrase
Convert the feature vector into 1 to 4 segments with explicit timing and edge operators.
Pink.short
→ Blue.medium
→ Yellow.open
3 · Read back
Infer utterance family
A human or model sees a likely class: check-in, request, explanation, state report, landing.
Trace family:
relational + informative + open
≈ explanation request
Example phrase families
These are not word-for-word translations. They are readable utterance families that can be learned.
Check-in family
Relational opening
Used for: “How are you?”, “Are you okay?”, “You there?”
Pink.short → Red.medium → Yellow.open
State report family
Subject + field condition
Used for: “I’m fine.”, “I’m home.”, “I’m okay.”
Red.short → Green.long
Transit family
Underway in system space
Used for: “I’m in the car.”, “I’m on the train.”, “I’m on my way.”
Red.short → Purple.long
Request family
Activation without closure
Used for: “I need help.”, “Can you do this?”, “I want this.”
Red.medium → Orange.medium[pulse] → Yellow.open
Explanation family
Structured information
Used for: “Here’s what happened.”, “This means…”, “Let me explain.”
Blue.long → Green.short
Selection family
Choice inside a structured field
Used for: “I’m buying fish.”, “I choose this one.”
Red.short → Blue.medium → Orange.short → Green.short
Minimal v1 specification
A first implementation can stay compact. The goal is not maximal symbolic coverage, but fast legibility.
Notation
| Code | Meaning |
|---|---|
| P.s | Pink short |
| R.m | Red medium |
| O.m[p] | Orange medium with pulse |
| Y.o | Yellow open ending |
| B.l | Blue long |
| G.l | Green long landing state |
| V.l | Purple long system field |
Render rule
Segment = ⟨ hue, length, weight, edge, pulse, confidence ⟩
Trace = [ S₁, S₂, S₃, ... Sₙ ]
Example:
“How are you?”
= [ P.s, R.m, Y.o ]
“I’m in the car.”
= [ R.s, V.l ]
“I need help.”
= [ R.m, O.m[p], Y.o ]
Closing line
The sidecar should say less than language, but reveal the shape of language faster.