A.1:10 - Rationale
A.1 prevents category errors by separating individuation, constructive part-whole recognition, acting eligibility, and claim-bearing. U.Entity gives the minimal referenceable object. U.Holon adds six separately recoverable constructive components: candidate, constituents, part-relation occurrences and assembly, reidentification, a composition-grounded whole-level characteristic, and compatible possible participation in a governed larger assembly. U.System adds acting eligibility. U.Episteme adds claim-bearing structure without agentivity. U.Work and U.Discipline are holon-kind examples only through their subject patterns; BoundedModelUseStructure is selected U.Structure, not another holon kind.
The recognition base cannot depend on a prior context object without recursion. It begins with the candidate and world-side construction facts. Public-kind admission is the separate one-time E.24.UK decision. Classification work evaluates what its available inputs establish about those facts and may return true, false, or unknown; the evaluation and its record remain distinct from the construction and identity being judged.
This also prevents ontology duplication. A theory under concern, a theory description, a publication of that description, and the system that edits the publication can all be named without turning the filling of one participant slot into a new kind. Architecture likewise starts from the holon recognized under an admitted kind whose selected structures matter; diagrams and structure descriptions remain epistemes.
The constructional stance is conservative: FPF avoids unrestricted composition and uses A.14 and C.13 before a part-whole claim is relied on for another claim or work occurrence; B.3.5 is added only when a named assurance use elects its profile. This keeps holonic thinking useful without letting every collection, expression, graph, selected structure, or source label become a holon.