Extrinsic site defensibility and landscape-based archaeological inference: An example from the Northwest Coast

• I generalize Martindale and Supernant’s (2009) extrinsic defensibility indices to a raster landscape.
• Raster results for the Gulf of Georgia region are comparable to those developed by Martindale and Supernant (2009).
• Site locations in the Gulf of Georgia region are significantly more defensible than the landscape as a whole.
• Trench embankment sites are often in highly defensible places on the landscape.

People make decisions in the context of their physical and social environments. Therefore, when inferring the choices that people may have made in the past, archaeologists should consider—to the extent possible—the environmental context(s) of decision making. In this paper, I attempt to build stronger inferences about the nature of defensive decision-making by characterizing the defensibility of a given landscape and treating it as a population from which a sample of archaeological sites may be considered. I develop a spatial defensibility index that may be calculated for any and all points on a raster landscape (a digital elevation model). I then calculate the defensibility of a large region in Gulf of Georgia and lower Fraser River valley of British Columbia, and assess the defensibility of a large sample of recorded pre- and post-contact archaeological sites in light of the baseline defensibility of the landscape. I find that while residential sites are generally built in more defensible places on the landscape, previously identified “defensive” sites (trench embankment sites) are not necessarily in unusually defensible places. These and similar methods ought to be employed whenever archaeologists attempt to infer defensive decision-making, and are essential for cross-cultural study of warfare and conflict.