CHRONOS alignment follow-on: cross-reference neuro-symbolic method-family taxonomy into vocabulary drafts (per sociosphere#498)

[mdheller]

Context

SocioProphet/socioprophet#498 tracks the overall KAIROS↔CHRONOS alignment thread. sociosphere/docs/integration/neurosymbolic-chronos-alignment.md's authority-plane table assigns Ontogenesis the "Semantic vocabulary draft" role. Its "Definition of done" requires a downstream follow-on item here. This is that item.

What already exists in this repo

Ontogenesis already has two CHRONOS-adjacent vocabulary tranches, but for an earlier/different CHRONOS lane than the ASU/neuro-symbolic one:

• docs/corpus-event-semantics.md (v0.1) — "the first Ontogenesis carrier vocabulary for the Watson/Cyc/Semantic-Web/CHRONOS corpus substrate merged in SocioProphet/sociosphere#334 and coordinated through SocioProphet/sociosphere#335." Defines ces:CorpusEvidenceBundle, ces:EventInstance, ces:CausalRelationCandidate, etc. for the loop EvidenceBundle -> EventInstance -> ActionProposal -> PolicyDecision -> AuditEvent.
• docs/symbolic-regression-vocabulary.md (v0.2) — defines EquationCandidate -> SRAssertionProposal -> SRAssertion for PROMETHEUS/symbolic-regression, with vocabVersion/vocabularyPromotionState (draft/stabilizing/canonical) and an explicit mention of "any CHRONOS risk flags that force human review" as a TODO for the AgentPlane sprint gate policy.

Neither document yet uses the specific method-family taxonomy from neurosymbolic-chronos-alignment.md (Logic Review, Kautz/NSR, LTN, LNN, NeurASP, SATNet, dILP, DON/RRN, DSR/DSP) or the alignment doc's carrier-boundary field list (grounding status, validation status, explanation trace reference, owning authority plane, non-authority declaration, replay reference).

The concrete gap

Ontogenesis owns "vocabulary draft" for two active CHRONOS-adjacent lanes already, but the ASU-derived method-family classification and the newer carrier-boundary field list aren't cross-referenced into either corpus-event-semantics.md's classes or symbolic-regression-vocabulary.md's SRAssertionProposal. Concretely: an EquationCandidate produced by a dILP-style rule-learning method and one produced by ordinary symbolic regression currently look identical in this vocabulary — there's no field carrying which method family produced a candidate, so downstream consumers can't apply the alignment doc's per-method admissible/forbidden-use rules (e.g. "a learned rule is promoted as canonical schema" is a named negative rule in the alignment doc but nothing in SRAssertionProposal currently declares whether a candidate came from a method family that rule applies to).

Ask

Look at whether corpus-event-semantics.md's carrier classes and/or symbolic-regression-vocabulary.md's SRAssertionProposal should adopt (or cross-reference) the method-family vocabulary and required carrier-boundary fields from sociosphere/docs/integration/neurosymbolic-chronos-alignment.md. This is a request to evaluate the connection, not a mandate to merge the two vocabulary tranches or to adopt any specific field set.

Non-goals

• No canonical-schema authority moves into sociosphere; symbolic-regression-vocabulary.md already states SourceOS-spec may receive canonical schemas only after this draft stabilizes — that boundary is unchanged.
• This is not a request to make WebProtege or any specific review surface mandatory.
• This does not ask Ontogenesis to take on AgentPlane's evidence/replay ownership or Policy Fabric's admission ownership — only to consider whether its own vocabulary drafts should be able to express which neuro-symbolic method family produced a candidate.

Refs: SocioProphet/socioprophet#498, sociosphere/docs/integration/neurosymbolic-chronos-alignment.md, docs/corpus-event-semantics.md, docs/symbolic-regression-vocabulary.md (this repo).