Natural-product research often treats botanical identity as though it were a single property that remains stable from a named species to the material placed in a biological assay. In practice, evidence may refer to a taxon, a defined plant part, a processed extract, a chromatographic fraction, an analytical feature, a candidate structure, or an isolated compound, and these objects are not evidentially interchangeable. This original classification article examines chemical identity as a provenance problem and proposes a taxonomy that separates material state from the confidence and scope of chemical identification. The synthesis distinguishes taxonomic/source identity, plant-part and material-form provenance, extract identity, fraction identity, and isolated-material identity, while treating chemical profiles, mass-spectrometric features, class assignments, candidate structures, and confirmed structures as separate analytical evidence states. A central principle is that biological evidence attaches first to the material actually tested; transfer to another pharmacognostic level requires an explicit bridge justified by shared provenance and appropriate analytical evidence. This prevents species names from functioning as proxies for extract composition, active fractions from being treated as proof of a single causal constituent, and computational annotations from being represented as established structures. The taxonomy is intended as an evidence-organizing framework rather than a validated scoring system or regulatory standard. Its usefulness will depend on reproducible classification, adequate reporting, orthogonal authentication where needed, and prospective testing of whether provenance-aware interpretation reduces unsupported evidence transfer.
