Collective-First Causal Closure in a Finite Heritable Ecology: Exact Organization Search, Heritable Amplification, and Ecological Reversal

Can a collective whose continuation depends on interactions among components with no assigned independent viability target arise and be amplified under a specified selection process? We analyze a three-cell Boolean ecology at two levels. An exhaustive developmental census of 98,304 organizations finds 192 surviving every four-tick environmental history and one minimum-complexity surviving organization; its best uncoupled comparator survives 160 of 256 histories. Removing either required incoming peer dependency reduces the selected controller's survival from 256/256 to 160/256. Its ordinary continuation dynamics admit a five-class exact quotient, refining to 17 classes under a declared intervention family. In a separate exhaustive genome space of 2,654,208 heritable organizations, 2,304 (0.0868%) survive all histories, and six attain the minimum cost signature. A preregistered, repository-reported 64-lineage comparison finds that 46 selected lineages, but no neutral lineage, attain at least 25% full-survival occupancy. After changing the resource law to eliminate the need for delayed distributed information, mean peer-edge count declines from 1.569 to 0.214; a sham continuation stays near 1.573. A new code-disjoint reproduction confirms the complete developmental and genotype censuses. The original selection trajectory study has internal reexecution evidence, but its full independent source-faithful rerun remains outstanding. The contribution is an explicitly conditional finite proof-of-mechanism, not a demonstration of abiogenesis, subjective experience, or morality.