Controlled vocabulary for interpretive governance

InferensLab uses vocabulary as a stabilization tool. Terms are not decorative. They are meant to reduce ambiguity, normalize distinctions, and keep doctrine interpretable across pages, files, and machine surfaces.

The glossary is short on purpose. It is a control surface, not an encyclopedic dictionary.

Core terms

• Interpretation: meaning assigned to an AI output in context.
• Distortion: gap between plausible interpretation and correct or authorized interpretation.
• Conflation: merging distinct concepts into a convenient narrative.
• Drift: uncontrolled variation across equivalent cases.
• Machine-first surface: machine-readable entrypoint designed to reduce ambiguity.
• Evidence artifact: minimal trace that makes an interpretation verifiable.
• Response legitimacy: condition under which an answer may be produced at all.
• Anti-plausibility: rule forbidding completion of absent facts by coherence alone.
• Authority boundary: line between what may be deduced and what would require inference or permission.
• Canonical silence: legitimate absence of answer when the surface does not authorize completion.
• Interpretive debt: silent accumulation of ambiguous, weak, or misaligned public signals.
• External Authority Control (EAC): governance layer that declares which external authorities are canonically admissible.

How to use the glossary

• Use the glossary to normalize meaning across pages and files.
• Treat glossary terms as doctrinal anchors, not as marketing labels.
• When a term appears in a note, prefer the controlled definition over external paraphrase.

Companion machine files

/.well-known/ai-glossary.json carries the controlled term list. evidence-taxonomy.json and doctrine-taxonomy.json provide adjacent controlled structures.