Specialized-metabolite profiles are widely used to compare plant taxa, support identification, and interpret evolutionary differentiation, yet the chemical profile measured from a specimen is not necessarily a fixed expression of species identity. It is an observation produced jointly by inherited metabolic capacity, tissue and cell specialization, developmental and inducible state, environmental exposure, sampling time, and the analytical workflow used to recover and annotate chemistry. This article develops an original theoretical model for chemotaxonomic inference under such metabolic plasticity. The model separates independent phylogenetic evidence from biological chemical state and from the analytically observed metabolite representation, thereby preventing chemical similarity from being interpreted automatically as common ancestry or chemical difference as taxonomic separation. It further proposes that chemical traits should be interpreted according to whether their taxonomic information is stable across relevant contexts, conditional on explicitly matched contexts, or potentially misleading because biological or analytical states are mismatched. The framework retains convergence, rapid lineage divergence, inducible metabolism, seasonal change, tissue compartmentation, and incomplete chemical identification as competing explanations rather than treating them as residual noise. For classification and authentication, the model favors concordance among chemistry, independent identity evidence, provenance, and context while treating discordance as an object of investigation. The framework is theoretical rather than prospectively validated; it supplies no universal marker thresholds, weighting coefficients, or diagnostic performance claims. Its principal value is to make the conditions under which chemical evidence can legitimately support taxonomic inference explicit and experimentally testable.
