Defining Support-Based Classifications from Aggregated Relational Records

A support attribute can be explicitly defined while remaining unresolved by an aggregated record. This paper examines that distinction in finite models of proposals supported by persons partitioned into groups. It separates attribute representation, definition relative to a language, evidential assignment, and institutional use. After reconstructing the elementary symmetry test for definability, it gives an explicit formula for a fixed threshold of distinct supporters. An exact group-incidence record preserves which groups contain support while omitting individual multiplicity. Under known finite capacities and otherwise unconstrained support relations, the paper proves that every person count between the number of occupied groups and their combined capacity is attainable. This yields sharp threshold judgments and an exact condition for recovery across the whole record format. A fictional dossier distinguishes person counts, group diversity, eligibility, and decisions under uncertainty. Counterexamples address overlapping memberships, shared endorsement constraints, incomplete negative records, and category revision. Verified antecedents in model theory, incomplete databases, numerical aggregation, and conceptual revision locate the inquiry without a claim of priority. The institutional discussion states a conditional premise for answerability and keeps authority distinct from definability. Open questions concern targeted record refinement, joint uncertainty, support meaning, and evaluation in actual practices.