Complete SOUL-WP-0009: resolve RQ4 and publication pathway, deliver T02-T05

- T01 (RQ4): retain "Soul" - resolved with operator; Terminology.md,
  concepts/soul.md, OpenQuestions.md, Claims.md, paper/draft.md updated.
- T02: lexicographic continuity-weighting procedure in concepts/identity.md,
  validated against TE-02/03/04/08.
- T03: testable Being-vs-Social-Being boundary (three-part recognition test)
  in concepts/recognition.md and concepts/being.md, validated against
  TE-05/09.
- T04: research/metrics-scoping-note.md - per-axis measurability scoping.
- T05: research/law-governance/moral-legal-personhood-charter.md - bounded
  companion-line charter, explicitly not the mapping itself.
- T06: publication pathway resolved with operator - internal circulation
  only, as paper-one for coulomb-research, following the same staged
  trajectory as coulomb-research's own paper-zero.

Marks SOUL-WP-0009 finished. Opens SOUL-WP-0010 (proposed, not started) to
carry forward the two multi-step residuals from T04/T05.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

Assistant: claude-code
Assistant-Model: sonnet
Assistant-Process: 136745@bnt-lap001
Assistant-Session: 2e334b58-ed31-48ed-8156-d09f8ed0781f

concepts/recognition.md

@@ -40,6 +40,44 @@ Recognition alone does not create systemic individuality (RN-001 Finding 10).
 > distinguish it, address it, expect continuity from it, maintain models of it,
 > and enter persistent relationships with it.
 
+## Testable boundary: Being vs. Social Being (SOUL-WP-0009-T03, resolved 2026-09-28)
+
+**Boundary condition:** Social Beinghood begins, for a given recognizer,
+the moment that recognizer holds a **persistent, addressable model** of the
+candidate across more than one interaction episode — not merely a one-off
+epistemic classification. Concretely, three things must hold simultaneously
+on at least one side of the relation:
+
+1. **Intentional recognition** — the recognizer treats the candidate as an
+   agent with goals/beliefs, not just a classified pattern.
+2. **Addressability** — the recognizer can specifically target the same
+   candidate again (name, ID, channel), not just re-classify a similar one.
+3. **Model persistence across episodes** — the recognizer's model of the
+   candidate carries over from one interaction to the next (expectations,
+   history, revisable beliefs about it), rather than being recomputed from
+   scratch each time.
+
+**Reciprocity is explicitly not required to start it** (TE-09): the three
+conditions above need to hold for only *one* party. A single Being's
+persistent, addressable, intentional model of an otherwise-unrecognizing
+system is sufficient to open Social Beinghood on that Being's side of the
+relation — this is what makes the asymmetric case in `paper/draft.md` §1
+and §5 well-formed rather than circular.
+
+**What this rules out:** a system that is epistemically classified fresh
+each time (no model persistence) is not Socially a Being to that recognizer,
+however accurately it is classified — this is the stateless-agent case
+(TE-01) restated from the recognizer's side. A system recognized once and
+never modeled again likewise does not cross the boundary.
+
+**Nested/collective test (TE-05):** apply the same three-part test
+independently at each candidate level. A collective can be Socially a Being
+at the collective level (some recognizer holds a persistent, addressable,
+intentional model of *the collective as one thing*) without that implying
+or requiring Social Beinghood at the member level, and vice versa — the
+levels are tested separately, not inherited automatically in either
+direction.
+
 ## TE-09 (summary)
 
 Unrecognized coherent intelligence can be an autonomous individual without