Functional Audit and Candidate-Coverage Diagnostics for Connection Dynamics

Abstract

Under frozen criteria and in artificial environments, some relation combinations pass an existence gate, a task-synergy gate, and a limited set of control gates; offering identity opportunities to historically selected targets can raise retention rates on some seeds. This supports a coverage-and-opportunity diagnosis in specific models. It does not support blind discovery, full strong generalization, reflexive organization, phenomenal consciousness, or physical non-propagation. v0.5 corrects and refines Section 5.7 after a candidate-by-candidate cross table of the frozen specificity gate against a same-target pool test: among 564 audited active same-target pairs the six named controls discriminate (133 below the floor) while the matched-random, shuffle and synergy margins and the pool test discriminate weakly, and the pool test passes all 564 including the 139 that fail the frozen gate. This replaces the v0.4 wording that treated all frozen controls as weakly discriminating. Section 5.8 (descriptive chain-specific comparison, 8 chain-strong records) and Section 6.1 (H1 restated as two layers) are carried over with adjusted wording. The reproduction bundle is extended (derived tables plus a 4,990-row pair-level table; the 2.5 GB raw landscape file is not included). Post hoc analysis of archived data; no new learner was run. Unpublished discussion draft; not peer reviewed.