Computationalism as Structural Non-Explanation_ Why Computation-First Accounts of Mind Cannot Converge to Ontology

Abstract Computation-first accounts of mind—commonly grouped under computationalism and functionalism—are often treated not merely as modeling frameworks but as ontological explanations of experience. This paper audits that ontological claim at the level prior to empirical dispute or engineering progress. The question addressed is not whether computational models are useful, predictive, or indispensable, but whether computation, when taken as explanatorily primary, can converge toward a determinate ontology of mind. Using a strictly structural standard, the paper evaluates whether computation-first explanations satisfy three necessary conditions for ontological adequacy: intrinsic causal grounding, eliminability of competing ontologies in principle, and contraction of admissible accounts under refinement. It is shown that computation, by its nature, admits unlimited realizations under re-encoding and abstraction, depends on observer-relative interpretive mappings, and lacks a fixed commitment surface that would permit eliminative discrimination. As a result, refinement expands rather than contracts the hypothesis space. The failure identified is not empirical, technological, or provisional. It follows from the explanatory form of computation itself. Computationalism therefore cannot function as an ontology of mind, though it remains admissible and indispensable as a descriptive and predictive framework. The paper concludes by reclassifying computation-first theories accordingly and distinguishing them from mechanistic, biophysical, and constraint-based approaches that survive the same audit by fixing causal commitments and admitting convergence.