A General Theory of Selection

Selection from astronomically large candidate spaces recurs across domains as disparate as subatomic physics, molecular biology, and corporate finance. We propose the General Selection Principle (GSP): the observed configuration is selected by the conjunction of two independently motivated filters—admissibility (structural/arithmetic consistency) and viability (a dynamical threshold on generative-to-dissipative rates). The viability filter requires the rate ratio to exceed the Information Profit Threshold $ = 1 + /2 (2 ) 1.1309$, derived from first principles in . Together, the filters produce asymptotically sparse but non-empty survivor sets. We document two independent quantitative signatures of the GSP. The IPT threshold recurs across five empirical domains — ecology and economics (prior derivation ), nuclear magic numbers, prebiotic chemistry, corporate survival, and microbial metabolism (FBA, r=0.971, p<0.001) — together with one model-internal symbolic-communication result (natural language alphabet size). The algebraic substrate Q recurs in the Standard Model parameter spectrum , the Zamolodchikov E8 integrable QFT mass ratios , ADE Toda field theories, and SU(2)_k WZW quantum dimensions. The E7 Toda theory (h=18, 18 120) provides a clean falsifier: all six mass ratios lie provably outside Q, with the exclusion Lean-certified . Evidence grades range from [T] (theorem-level, machine-certified) through [B] (empirical, replication pending) to [B^-] (empirical, internal dataset); [Br] denotes bridge claims conditional on companion-paper premises; a domain-by-domain evidence map appears in Table . The co-occurrence of these two independent signatures supports the hypothesis that a single mathematical mechanism underlies selection across discrete combinatorial systems; see also the companion SRRG paper for a proposed mechanism.