Placement study: accuracy-of-meaning + externalization into DESIGN.md

[briansrls]

Analysis only — the deliverable is the plan document; no authority is edited. gunbc.design_document, gunbc.recurring_failure_mode, and both projections (DESIGN.md, docs/design-ledgers.md) are untouched: the operator's supervised edit performs the actual insertion, starting from the copy this plan drafts (marked NOT APPLIED). The corpus-wide violation audit is carved out to the operator.

Delivers docs/plans/meaning-externalization-design-placement.md:

• Placement verdicts (each concept gets one home, as a consequence of its section's existing law):
  • Accuracy of meaning → §3, one paragraph: the meaning fork (one name, two concepts) is the exact dual of nicknaming (two names, one concept); the single-authority law is the name↔meaning bijection read in both directions, extended to every layer that names — product names included.
  • Externalization → §5 named trap, single home: the absorbing fallback with a counterparty as the absorber. Stating it in §2 as well is refuted — §2's later-fixer sentence already stands, and the serial charter forbids restatement; the new content is the refusal discipline (absorb-and-reserve or name-and-price), which is §5's genre.
  • Quality floor → §4b, one paragraph: the ladder applied to counterparty-facing promises — a claim with no executing consequence is §4b's permanently-green decoration sold outward; named degradation is §4b(3) verbatim (deviation declared, silence refused).
  • Ledger: meaning_fork and externalized_degradation both belong as new gunbc.recurring_failure_mode rows, with recognition rules drafted; rungs stated honestly as review-diligence vocabulary (executing evidence today is private: strategy.pricing_decomposition + its witness).
• Literature census: exact insertion coordinates by symbol (section_3_blocks / section_4b_blocks / section_5_blocks, roster append), what regenerates, what is untouched (gunbc.rung_drop, gunbc.design_ledgers), and why no §2 edit / no lens / no rung-drop row.
• Downstream census: the derivable meaning-fork lens (dual of gunbc.bare_name_fork_lens), the review probes each row arms, which public constructions come under the higher bar, and the private evidence + five-gap audit as specimen backlog, with the honesty boundary stated.

The second file is the plan's HandAuthoredDocBind row in gunbc.doc_graph_roots — the registration route hand-authored plan docs land by (precedent: roadmap-launch-mvp-plan), binding the doc to its exact insertion targets with a dissolution on the supervised edit landing. Without it the doc is an unbound orphan (§6 experimental residue), so it is part of landing the doc, not scope beyond the brief.

🤖 Generated with Claude Code

https://claude.ai/code/session_019moVbKTZcEvxkXb1VjK9F8

[gunbai-bot]

Superseded: consolidating onto #9769 as the single plan vehicle per the operator's ROOT-N program (ROOT-0). This PR's review 'REQUEST CHANGES' findings (one-file boundary violation via the doc_graph_roots row; §4b conflation; witness-coverage overstatement) are being applied as rewrites on #9769. — sent from jolly-swift-749