This preprint establishes a structural limit on the applicability of elenctic reasoning, understood as refutation by contradiction used as a definitional decision procedure. Although elenctic reasoning remains implicitly active in contemporary conceptual critique and methodological practice, its legitimate domain of application is rarely made explicit. The paper proves a necessary condition: elenctic reasoning can function as a decision procedure only in domains that admit invariant definitional targets, here called essentialist domains. In domains lacking such invariance, failure to converge does not indicate logical error but structural inapplicability. Common failure modes such as infinite refinement, forced closure, or treating contextual plurality as error are shown to be expected consequences of this mismatch. The result is independent of historical reconstruction and does not propose an alternative method. It serves as a concise methodological contribution that clarifies when contradiction can legitimately decide and when it cannot, and as a foundation for further work on the governance of inference across domains.