Weksa/schemas/language-project-state.md
2026-05-16 03:43:43 +01:00

1.1 KiB

Language Project State

This is the intended future state shape for an individual generated language. It is not executable yet.

Ownership Families

  • metadata: language name, world/culture context, authorship, version.
  • ontology: concept inventory, semantic relations, splits, collapses, salience.
  • interlingua: source meaning representation accepted by the renderer.
  • grammar: syntax, agreement, evidentiality, aspect, valence, required categories.
  • morphology: roots, affixes, compounding, derivation, inflection.
  • phonology: phonemes, syllable shapes, phonotactics, prosody, romanization.
  • diachrony: proto-forms, sound changes, semantic drift, strata, registers.
  • lexicon: generated forms, glosses, etymologies, usage constraints.
  • corpus: example sentences, analyses, translations, provenance.
  • critique: model or human findings about consistency and plausibility.

Boundary Rule

The language project should not store raw model completions as truth. A model completion can be evidence, suggestion, or critique. Accepted language state must be normalized into the owned families above.