research-document RJ-VE-ITTEN-001
Itten Color Contrasts — Research Cycle 1
Research cycle 1
Question
Which of Itten's seven contrasts remain scientifically defensible, what mechanisms and boundaries replace his explanations, and which claims should become engineering rules?
Iterations
Pass 1 — decomposition
Separated Itten's taxonomy into stimulus dimensions, perceptual effects, semantic associations, pair relations, and compositional judgments. This invalidated the hidden assumption that all seven contrasts are comparable mechanisms.
Pass 2 — mechanism and falsification
Searched for opponent processing, appearance models, brightness/color induction, categorization, aging, color-vision deficiency, HDR/ambient behavior, and saliency. Explicit counterevidence included induction-to-assimilation reversals, cross-language departures from Hering-style decompositions, and failure of increased saturation to improve overall CVD search performance.
Pass 3 — engineering translation
Rejected universal palette recipes. Replaced them with contextual specifications, rendered-state testing, retinal-scale constraints, redundant encodings, device characterization, and behavioral validation.
Revisions
| Revision | Supersedes | Reason |
|---|---|---|
| Warm/cool demoted to association | sensory “temperature” axis | measurable mappings do not establish direct thermal perception or universality |
| Complement split four ways | single complementary relation | additive, opponent, afterimage, and subtractive complements can diverge |
| Extension retired as color law | fixed area harmony | salience and balance depend on many interacting features and task |
| Simultaneous broadened | retinal lateral inhibition only | early, cortical, grouping, multiscale, and inference evidence coexist |
| Saturation renamed by attribute | one intensity dimension | colorfulness, chroma, and saturation have distinct definitions |
Saturation decision
Further unsystematic source collection was no longer changing the seven dispositions. The next meaningful gains require formal systematic searches, effect-size extraction, and new experiments, especially on area balance, spatial color models in UI, HDR adaptation, and underrepresented observer groups.