The first phase of this project consists of inventorying artifact collections (consisting mainly of whole ceramic vessels) at several museums around the country. These collections represent the primary data set for investigating pre-contact art, ritual, and social interaction in our study area. Our inventory work provides the information we need to determine when the vessels were made (chronology), which communities produced the vessels (cultural affiliation), and what motifs and themes are reflected by vessel decorations and related artworks (iconography). The results of these analyses provide a framework for identifying production rules and principles, inferring associated conceptual frameworks, and tracing the extension of those frameworks to community identity representations and social interactions.

Recent studies of ceramic vessels in the central Mississippi River Valley and in the Ouachita Mountain/Gulf Coastal Plain regions offer new approaches to guide our investigations. Robert Mainfort generated a series of studies demonstrating the value of multivariate statistical analyses that include vessel form attributes, such as rim shape and angle, in addition to decorative motifs to validate regional variations associated with chronology and cultural affiliation in the central Mississippi Valley (Mainfort 1999, 2003, 2005; see also McNutt 2008). In a recently completed M.A. thesis, Walker (2008) employed some of Mainfort’s concepts to refine our understanding of the multiple ceramic-making traditions represented in the Carden Bottoms collections.

In a study of Caddo ceramics from southwest Arkansas, Ann Early (n.d.) demonstrated that local potters employed design rules to decorate their vessels with such specificity that it is possible to differentiate ceramics made, for example, in the Ouachita River Valley from those made in the Red River Valley. These rules represent design grammars reflecting local pottery-making traditions keyed to specific community identities. Early also discerned in the design grammars of her study assemblages a set of underlying principles similar to those Sabo (1998) identified in his analysis of historic Caddo kinship terminologies, mythology, and household spatial organization. The Caddo conceptual framework, centered on the maintenance of hierarchical relationships organized with reference to principles of seniority (e.g., parent/child) and relative strength, represents a fundamentally different set of ideas than those expressed in the Siouan landscape model, identified in our rock art study, reflecting non-hierarchical principles of complementary opposition and reciprocity. In sum, these examples demonstrate that reconstructions of art production rules and design grammars provide a useful approach to understanding how ancient communities expressed their identities at particular moments in time and linked those identities to broader conceptual frameworks.

Levels of confidence attributed to determinations of chronology, cultural affiliation, and iconography based on analysis of existing collections vary from poor to better in relation to the quality of contextual information available for the assemblages under study. Rock art, fixed as it is within local community landscapes, provides the most secure information for such studies. If the ceramic vessel collections for our study area had a comparable level of provenience information, we would need to go no further in our study than to analyze those collections according to our existing protocols. However, the majority of ceramic assemblages from our study area have no excavation provenience; most vessels are attributed to a site or locality based on the claims of the looters who dug them up. What confidence can we place in such claims?

First, it is possible to independently verify site/locality provenience claims. Local clays used in the production of ceramic vessels are distinctive in visual appearance and texture, and vessels made in the Carden Bottoms can usually be sorted from vessels found at the same sites but made elsewhere. Comparison of the application of design grammars and subtle variations in vessel shape can also be used to separate local copies from similar trade pieces, as demonstrated in a recent study by Early, Walker, and Sabo (2008).

Since the majority of vessels in museum collections came from cemeteries, we can also examine other site contexts such as houses, middens, storage/trash pits, and work areas from which artifact, faunal and floral remains, and radiocarbon samples can be collected. Excavations of such features at sites associated with the looted cemeteries will corroborate the site attributions of specimens in existing collections, extend our understanding of the variety of archeological contexts in which such specimens occur, clarify distinctions among ceramic assemblages from mortuary and other domestic contexts, and expand our understanding of the general structure of archeological sites in the Carden Bottoms locality (e.g., Stewart-Abernathy 1994; see also http://www.uark.edu/campus-resources/archinfo/atucarden.html).

With such prospects for validating the archeological contexts of existing museum collections via excavation of associated site areas, what remains to be incorporated in our research design is a method for efficiently targeting specific archeological features at those sites. How can we “zoom in” on features representing the domestic and ritual contexts we seek to explore? This is where our remote sensing and Geographical Information Systems applications come into play.