A model for assigning temper provenance to archaeological ceramics with case studies from the American Southwest

Abstract

Well-designed provenance studies form the basis from which questions of human economy and behavior are addressed. Pottery is often the subject of such studies, requiring geological and archaeological evidence to establish patterns of ceramic economy. A generalized theoretical and methodological framework for provenance studies is presented, followed by specific considerations for ceramic provenance studies. Four main sources of variation affect pottery composition: geological distribution of resources, geological resource variability, differential economic factors affecting resource use, and technological manipulation of materials. Post depositional alteration is also considered. This ceramic provenance model provides explicit guidelines for the assessment of geological aspects of provenance, since geological resource availability affects acquisition by humans and thus archaeological research designs, in which interdependent geological and archaeological scalar factors must be balanced against budgets. Two case studies illustrate the model. The first is of sand-tempered pottery from the Tonto Basin, Arizona, where the bedrock geology is highly variable giving rise to geographically unique sands. Zones with similar sand compositions are modeled using actualistic petrofacies, the Gazzi-Dickinson point-counting technique, and multivariate statistics. Methods used to create a petrofacies model are detailed, as is the model's application to sand tempered utilitarian sherds from three Tonto Basin project areas. Data analysis reveals strong temporal and spatial ceramic production and consumption patterns. The second is of crushed-schist-tempered Hohokam pottery. Crushed schist was often used to temper pre-Classic Hohokam plain ware pottery in central Arizona's middle Gila River valley. Systematic investigation of rocks from the Pinal Schist terrane in the middle Gila River valley was conducted to assess how many sources were exploited prehistorically, and whether schist or schist-tempered pottery were exchanged. Chemical analysis shows that the sources can be statistically discriminated from one another. Schist source data were compared to schist extracted from plain ware sherds and to unmodified pieces of schist recovered from two archaeological sites. Preliminary indications are that schist was derived from several sources. This model provides a flexible, archaeologically relevant framework for assessing temper provenance. Hopefully, archaeologists and petrologists alike will use it to define ceramic provenance research problems and communicate effective solutions to one another.