EMPIRICAL REGISTER: 72 FALSIFIABLE PREDICTIONS

Whether a broad theoretical framework is empirically exposed cannot be settled by counting its predictions. A framework may publish sixty and remain untouched by all sixty failing, because none was load-bearing; may publish results already established in the literature, whose confirmation confirms the source and not the restatement; and may publish claims whose denial its own structure permits equally well, which are preferences rather than constraints. This paper develops a method that measures exposure directly. Each candidate prediction is passed through a six-step admission cascade and typed by four properties: whether the framework's structure forbids the denial, which named premises the prohibition rests on, what fails inside the framework if the prediction fails, and how many independent parties must agree before a falsifying result is established. The method is applied to the Trisductive Physical Seed, a substrate-monist arrangement of seven domain-level structures on a single actuation axiom. Seventy-seven candidates yield five standing predictions. Three further entries are structural necessities carrying no framework consequence; two are necessities belonging to general relativity and the Kerr geometry. The remaining sixty-seven are relocated with dispositions recorded and nothing discarded. No result carries a consequence larger than the retirement of one domain premise, though this is partly a property of how the register was assembled, since every candidate is seated within a domain; one failure branch carries less: a monopole attributable to substrate topology rather than to a failure of closure retires no domain at all and revises a substrate property whose reach is left open. No result reaches the framework's root, which is unreachable by a limitative theorem rather than by construction. Two of the five are settled by a single confirmed event in any adequately equipped laboratory, with no consortium in the falsification path.