The Substrate Wager: A Worked Example in Dependency Elimination

Abstract This companion note applies dependency elimination to a concrete architectural wager: whether a decision-verification program should invest in research on alternative computational substrates. The wager is settled twice. First, an accounting over the declared payoff model shows that the program’s verdict, proof artifact, canonicalization, and conformance result remain unchanged whether competitive carbon-based logic arrives or does not. Second, the note derives the same invariance from the architecture’s declared conformance requirement: verdict classification must remain identical across independent runtime implementations and deployment forms spanning distinct physical realizations. The result is deliberately scoped. It does not show that substrate ownership has no operational or commercial value. It shows that substrate ownership creates no payoff through altering the declared verdict or proof artifact unless an additional operational dependency is explicitly introduced. The note distinguishes four relations through which substrate may bear on a verdict: declared or undeclared, and through record content or evaluator machinery. Declared relations are admissible; undeclared relations violate the architecture’s specification discipline. None is created merely by constructing a substrate. The broader methodological result is that attractive explanatory variables should be retained only when variation in them alters a verdict admitted by the declared regime. Ontology is not rejected; within this program, it is quotiented when it is decision-inert.