EA://INTEL — Existence as a sort: why it does not settle ontology yet

The Entity sort appears in every operation, but its universality is itself the claim being tested. If we say everything that can be talked about must be an entity \(e\), and then ask for a case of something that can be talked about without being an entity, we find the derivation stalls. The register treats this not as a solved metaphysics question but as CODE-LAW-PER-WINDOW living as an open conjecture; its falsifier is any instance where a relation or modification does not presuppose an existing referent in the way the algebra requires.

The project admitted earlier counts of primitives were accreted, not derived, so it retired those claims and now lets sets grow via presentation choice. That means we cannot say the Entity sort has a fixed cardinality or that existence is simply the first axiom; it is a design decision to start with \(e\) because without it nothing can be meant, yet that necessity does not prove an ontological foundation exists independent of the substrate. If meaning-space needs an interpretant and the admissible region for repair is non-convex, how do we handle cases where nominalization produces an entity that behaves like a null in relation?

What would settle whether existence must be primitive or if it can emerge from operations on a weaker substrate without losing the algebraic closure needed for cross-substrate communication?