The paper develops a data analysis methodology for the exploration of the variance in chemical composition of ceramic artifacts due to their typological, locational and temporal characteristics. The methodology allows for the characterization of ceramics based on these characteristics and the sequential assignment of unclassified ceramics to a region, subregion or site, with further assignment to specific ware and chronological period. The procedure uses INNA data on the elemental composition of ceramics in order to derive a series of classification functions for each category within each level of the taxonomic hierarchy. The procedure is used to investigate a provenance problem involving Chalcolithic ceramics from the Zagros region of western Iran. The results show that the elemental composition of ceramics, when coupled with multivariate statistical techniques, represents a powerful means for distinguishing ceramic provenance.