Significance and context in GIS-based spatial archaeology: A case study from Southeastern North America
Introduction
One of the most critical questions any scientist must ask herself/himself is, “do my methods effectively test my hypotheses?” For spatial archaeological research, this requires appropriate statistical methods rooted in the known cultural, environmental, and landscape contexts. In the early years of computer-based spatial archaeology, Kintigh and Ammerman (1982) argued for spatial analyses guided by human decision-making more than the available methods. Their approach did not devalue computer-based analyses—in fact, it stressed that we produce spatial statistical results and not just “eyeball” patterns—but proposed that they must be chosen based on their ability to provide results related to our archaeological question or hypothesis. I have been particularly interested in two primary principles of their approach: 1) significance: our methods must be rigorous and provide significant results, and 2) context: our data and analyses must be contextualized in the relevant theory and archaeological data and cultural information known about the past people and their society.
Although GIS solves some of the problems faced in the early years of spatial archaeology, it also introduces new problems. It is easier to synthesize spatial and contextual data, and a multitude of spatial analysis methods are easily accessed by many archaeologists. As a result, GIS has the ability to alleviate many of the problems Kintigh and Ammerman brought to light. However, GIS programs do not have the ability to identify the proper spatial analytical methods for our data or the ability to teach us geographic principles and theory. They also cannot provide contextual cultural and archaeological data on their own. These problems still require the prioritization of human decision-making.
Thus, my goal for this paper is to provide a simplified example of how I have interpreted and applied—for both good and bad—the above principles in my own research. I will do this by describing the process through which I examined the settlement location choices of a small sample of Mississippian and Piedmont Village Tradition (PVT) settlements from the North Carolina and Georgia Piedmonts in order to explore why sociopolitical complexity arose and persisted where it did. I analyzed data collected in a GIS using discriminant function analysis, which compares the characteristics of datasets, determines whether they are significantly different, and, if so, which variables most distinguish them. I used this approach to approximate the process of selecting settlement locations outlined by Stone (1996) in his settlement ecology research. The results found here support many of the existing ecological hypotheses about why Mississippian chiefdoms arose where they did in this region, but contradict some of those related to economic interaction. To conclude, I offer my thoughts on my approach with regard to significance and context.
Section snippets
Methodological and theoretical considerations
Kintigh and Ammerman (1982) identified the primary problem in incorporating spatial analyses into archaeological research as being the loss of context that often occurs when employing quantitative analyses, especially those that are computer-based. They outlined a hypothetical situation in which an archaeologist observes a set of data, attempts a nearest-neighbor analysis by hand, recognizes the complexity of the issue, and finally uses a computer, which returns basic spatial patterning data
Methods
To reconstruct past landscapes, I used digital elevation models, USGS sediment maps, modern river locations, mid-20th-century wetland location data, and 18th-century trail maps (Mouzon, 1775, Myer, 1971). I assumed modern USGS sediment type identifications were appropriate for examining the past. Erosion rates have certainly changed over the last 800 years (Trimble, 1974); however, that does not necessarily alter sediment types. In fact, a sediment texture analysis pilot project at the 31Yd173
Results
When compared to the random locations within their region, Mississippian settlement sites in North Carolina were most distinguished by being located in larger floodplains, having larger wetlands within their catchments, having larger viewsheds, and having less well-drained sediments within their catchments (Table 2). The Mississippian settlements in Georgia were similar, being distinguished by larger floodplains, larger wetlands within their catchments, more western facing slopes, less well
Discussion
The results above agree with aforementioned hypotheses by Smith, 1978, Beck and Moore, 2002, and Meyers (1995) that productive resource areas may have been locations where sociopolitical complexity first developed (Fig. 3). It also parallels arguments for why complex Mississippian societies arose in the American Bottom (Milner, 1996, Milner, 1998, Pauketat, 2004). The diverse sediment types also hint at risk management being another factor in where Mississippian societies arose and persisted.
Conclusions
My goal was for this analysis to display how I attempt, through methodological and theoretical integration, to address the problems of significance and context in my own approach to spatial archaeology. With regard to significance, there are several considerations. I believe an important part of every spatial archaeological venture should be testing for autocorrelation. Without a comparison to the background landscape characteristics (Kvamme, 1990) or random distributions across a landscape (
Acknowledgments
The research discussed here was funded by a National Science Foundation grant (BCS-1430945). I would like to first thank Meghan Howey and Marieka Brouwer-Burg for the invitation to participate in this special issue and the SAA session that preceded it and for all of their hard work in organizing both. Substantive comments by reviewers greatly improved this work. Several undergraduate students at Wake Forest University have assisted in the data collection throughout this project, including
References (49)
An analysis of factors influencing sixteenth and seventeenth century Haudenosaunee (Iroquois) settlement locations
J. Anthropol. Archaeol.
(2010)- et al.
Provenance signatures of sandstone-mudstone suites determined using discriminant function analysis of major-element data
Chem. Geol.
(1988) Integrating pattern and context in spatial archaeology
J. Archaeol. Sci.
(1984)Variation in mississippian settlement patterns
Iroquoian landscapes: people, environments, and the GIS context
Woodland Subsistence-Settlement Patterns in the Great Bend Area
(1978)- et al.
The burke phase: a mississippian frontier in the North Carolina foothills
Southeast. Archaeol.
(2002) - et al.
Specific language impairment and grammatical morphology: a discriminant function analysis
J. Speech Lang. Hear. Res.
(1998) - et al.
Topographic and cultural influences on walking the Ridgeway in later prehistoric times
Reconstructing “total” paleo-landscapes for archaeological investigation: an example from the central Netherlands
J. Archaeol. Sci.
(2013)
Mississippian chiefdoms: how complex?
Annu. Rev. Anthropol.
Woodstock culture and the question of mississippian emergence
Am. Antiq.
The Siouan Project: Seasons I and II. Research Laboratories of Anthropology, Monograph Series No. 1
Sex determination by discriminant function analysis of crania
Am. J. Phys. Anthropol.
Stress and Warfare Among the Kayenta Anasazi of the Thirteenth Century A.D., in Fieldiana: Anthropology, n.s., 21. Field Museum of Natural History, Chicago
Models of prehistoric site location for the upper Illinois river valley
Ill. Archaeol.
Settlement as adaptation: variability in iroquois village site selection as inferred through GIS
Upgrading site-catchment analyses with the use of GIS: investigating the settlement patterns of horticulturalists
World Archaeol.
The settlement ecology of middle-range societies in the western North Carolina Piedmont, AD 1000–1600
N. C. Archaeol.
Refining our understanding of haudenosaunee settlement location choices
The ecology of shifting settlement strategies in the upper Yadkin River valley, AD 600–1600
Multiscalar settlement ecology study of Piedmont village tradition communities, A.D. 1000–1600
Southeast. Archaeol.
Exploring prehistoric tribal settlement ecology in the Southeast: a case study from the North Carolina Piedmont
North Am. Archaeol.
Cited by (0)
The special issue was handled by Meghan C.L. Howey and Marieka Brouwer Burg.