Repairing Confirmation A Structural Classification

This paper provides a structural analysis of attempts to “repair” confirmation when empirical or theoretical support appears to fail. It shows that most proposed repair strategies—appeals to auxiliary hypotheses, probabilistic softening, model enrichment, or interpretive reframing—do not restore admissibility but instead obscure standing failure. Within a closed admissibility framework, confirmation is not a graded or resilient relation. Either standing is preserved under admissible redescription, or it is not. When confirmation breaks, it cannot be repaired by adding structure, relaxing criteria, or shifting interpretive stance without violating boundary constraints. The paper classifies the full space of proposed confirmation-repair moves and demonstrates that each collapses into illicit scope transport, silent redescription, or post-hoc parameter injection. The analysis is eliminative. No new theory of confirmation is introduced, and no methodology for salvaging failed claims is offered. Instead, the paper establishes a boundary condition: confirmation failure is terminal with respect to the evaluated claim. Apparent repairs do not recover legitimacy; they merely defer recognition of inadmissibility. By exhaustively classifying and excluding confirmation-repair strategies, the paper dissolves debates over underdetermination, theory-ladenness, and evidential rescue. It closes these debates by showing that their persistence depends on treating confirmation as something that can be rehabilitated after standing has already been lost.