智名意
Visible sequence, Unicode normalization, grapheme handling, render profile, and public-symbol eligibility.
Implementation handoff
Static claim status from the public registry. Human review is required before claim widening.
Glyph object
Four-layer glyph object specification
Visible expression, visual structure, embedding evidence, and canonical output stay separate so interpretation remains bounded and reviewable.
Static glyph deconstruction interface.
The interface separates visible expression from intended concept. It is read-only and approximate, not a certified translation or hidden semantic authority.
Visible mechanics
Components, order, containment, relation cues, and unresolved target markers remain separate from meaning.
Evidence lanes
Visual, structural, semantic descriptor, and ontology-lane evidence can agree or conflict.
Bounded gloss
Candidate output with warnings, confidence, and public eligibility. Interpretation stays approximate until reviewed.
approximate interpretationunresolved targetambiguous modifierinsufficient evidenceprivate-use boundarynot a certified translation
Why one row is not enough.
A serious glyph system needs to know what was typed, how it normalized, how it segmented, what rendered shape was seen, which primitives were extracted, which candidates were retrieved, and whether public output is allowed.
Boundary statement
Unicode characters are not glyph images. Font-specific shapes do not become public semantic authority. Internal visual analysis may support interpretation, but visible public output must remain assigned Unicode characters or valid public sequences.
The four layers.
Surface layer
Unicode sequence, original input, normalized forms, grapheme clusters, rendered preview, font profile, and public-output policy.
Structure layer
Paths, primitives, strokes, radicals, containment, adjacency, symmetry, order, relation graph, and component roles.
Embedding layer
Separate vectors for visual similarity, structural graph, semantic descriptor, ontology projection, source family, and context.
Canonical layer
Ontology-validated expression, gloss, warnings, confidence, phase-lock status, human review result, and public-output eligibility.
Suggested diagram: four transparent stacked panels labeled Surface, Structure, Embedding, and Canonical.
Example object.
The object below separates public form from internal evidence and canonical interpretation. It is an example shape, not a claim that this exact record is deployed.
{
"glyphId": "iota.approx.unit.bridge",
"surface": {
"display": "ɪ≃1",
"normalization": "NFC",
"unicodePolicy": "public-symbol-only",
"publicOutputEligible": true
},
"structure": {
"sequenceRoles": ["iota-mark", "approximation-operator", "unit-anchor"],
"relations": ["approximate-equivalence", "identity-scale-reference"]
},
"vectors": {
"visual": "vector-ref:visual:001",
"structural": "vector-ref:structure:001",
"semantic": "vector-ref:semantic:001",
"ontology": "vector-ref:ontology:001"
},
"canonical": {
"bestGloss": "iota approximately one",
"canonicalExpression": "Approximate(Iota, Unit)",
"confidence": 0.68,
"status": "emerging",
"warnings": ["approximate interpretation", "not exact translation"]
}
}Validation checklist.
- Original input stored, normalization checked, and grapheme clusters segmented.
- Private-use, noncharacter, and variation-sequence policy checked before public output.
- Render profile recorded and primitives extracted or marked unavailable.
- Candidate concepts ranked, ontology checked, and confidence calibrated.
- Warnings, human review status, and public-output eligibility recorded.
Internal-only status.
Internal-only means the system may use the record for analysis, review, or model improvement, but should not present it as public semantic output.
Weak provenance
The source family or review state is not strong enough for public explanation.
Font overfit
The interpretation depends too much on one rendering profile or visual style.
Ontology conflict
The candidate gloss violates type, relation, or public-output rules.
How this connects to existing pages.
This specification operationalizes the Unicode Boundary and Communication pages. It also gives Developer Integration and IOTA-1 Walkthroughs a stable object shape for examples.
Do not claim.
Do not imply that internal SVG or visual decomposition creates new public Unicode meaning. Public output remains governed by assigned characters and valid public sequences.
Deep route polish
Four-layer glyph record narrative
The visual stack anchors the rule that visible expression, structural decomposition, embedding evidence, and canonical output must stay separate.
Written narrative
The glyph object specification prevents a visible mark from being collapsed into a single asserted meaning. It gives reviewers separate places for Unicode surface data, visual structure, retrieval lanes, ontology checks, warnings, and public-output eligibility.
Concrete example
A glyph may look similar to an existing symbol but fail ontology checks. The record can preserve visual similarity while refusing a public meaning claim.
| Focus | What to inspect |
|---|---|
| Surface layer | What was publicly visible or rendered. |
| Structure layer | How the shape decomposes into relations. |
| Canonical layer | The bounded output after evidence and warnings. |
Evidence note
Exact translation is not asserted. The record supports bounded glosses with uncertainty, warnings, and human review.
Internal reading path
/expression-concept-gap/
Expression-Concept Gap in Interpretable AI
Why AI systems must separate visible expression from inferred concept, especially in glyph systems, IOTA-1, and evidence-first semantic conversion.
/developer-integration/
Developer Integration Guide for Interpretable AI Architecture
How to integrate Teleodynamic AI pages, glyph interpretation objects, evidence payloads, and Protocol5-style converter diagnostics.
/evaluation-lab/
Evaluation Lab for Interpretable Systems
Metrics, gates, plots, human review worksheets, and audit tests for Teleodynamic AI and semantic glyph interpretation.