Species boundaries in insects are increasingly revised as genomic, morphological, ecological, geographic, behavioural, and life-stage evidence are combined at finer resolution. Such revisions are often portrayed as a conflict between taxonomic accuracy and nomenclatural stability. That framing is incomplete because instability can arise not from revision itself but from poorly governed transitions between observation, species-level interpretation, and nomenclatural action. This perspective develops a framework for analysing revision decisions under evidential discordance. It distinguishes evidence signals from taxonomic conclusions, identifies recurrent sources of conflict among mitochondrial, nuclear, morphological, geographic, ecological, and model-based evidence, and separates biological discordance from technical artefact. Contrasting insect cases show that additional evidence may justify either splitting or synonymizing, whereas algorithmic or marker-specific partitions should not be treated as automatic species decisions. The article therefore argues that stable taxonomy does not require resistance to change. It requires transparent rules for recording why boundaries are revised, which evidence was decisive, which alternatives remain plausible, and how decisions may be revisited. Nomenclatural continuity and biological accuracy are treated as distinct objectives that can sometimes converge but should not be conflated. The framework proposed here is conceptual rather than prospectively validated and does not prescribe universal genetic, morphological, geographic, or reproductive thresholds. Its purpose is to make conflict explicit and taxonomic change reconstructable.
