Network Analysis of Archaeological Data from Hunter-Gatherers: Methodological Problems and Potential Solutions

Abstract

Network analysis using hunter-gatherer archaeological data presents a number of unique challenges. At the forefront of these challenges are issues associated with the aggregation and fragmentation of archaeological data that influence the size, density and confidence in network models. These methodological challenges are unfortunate, as the diverse roles of social networks among hunter-gatherers have long been recognized within anthropological research. In order to enhance the research potential of networks constructed from hunter-gatherer archaeological data, this research highlights two data evaluation methods established in social science research to assess the stability of network structure. More specifically, this research constructs network models from the compositional analysis of ceramic artifacts recovered in the Kuril Islands of Northeast Asia and evaluates network centrality measures using bootstrap simulation and sensitivity analysis. Results of this research suggest that while archaeologists may never fully identify the "true" network of past relationships, network models that approximate "true" network structure can provide useful metrics in exploring the behavior of past hunter-gatherer populations. Overall, given the challenges associated with hunter-gatherer archaeological data, it is argued here that critically evaluating the structure and stability of network models is an essential first step in developing an archaeological network analysis that is relevant and informative to research on past small-scale societies.

Networks in Archaeology: Phenomena, Abstraction, Representation

“…However, they are not necessarily an end in themselves, and both the rationale for and form of visualization must be carefully considered if the resulting image is to achieve its aim and not simply end up as a so called "spaghetti monster"-a network so dense and complicated that it is extremely difficult to comprehend. Gjesfjeld;see also Mills, Clark et al 2013). …”

“…has framed this in terms of the multi-scalar ways in which materials intercede between people and even how things may interact with other things. We find in this volume, for example, the study of population growth, migration, and regional interaction within changing environmental conditions explored in the paper by Borck et al (2015); food exchange and settlement patterns simulated in the paper by Crabtree (2015); production and exchange of a variety of different kinds of raw materials, goods, and practices in those by Golitko and Feinman (2015), Gjesfjeld (2015), Graham and Weingart (2015), and Mol et al (2015); the diffusion of technological processes (Östborn and Gerding (2015)); and site interconnections, power, and intervisibility, as studied by . This kind of abstraction is aided by the archaeological record, since sites or assemblages of material culture form natural nodes, and seeking to focus on the dynamics by which they come to be characterized the way they are is an extremely fruitful line of analysis, yielding new interpretations and posing new research questions, as all these papers demonstrate in different ways.…”

Resilience and the population history of the Kuril Islands, Northwest Pacific: A study in complex human ecodynamics

“…As detailed elsewhere (Phillips and Speakman, 2009; Phillips, 2011; Gjesfjeld and Phillips, 2013; Gjesfjeld, 2015), evidence for changes in Epi-Jomon and Okhotsk network structures are based on a series of geochemical sourcing studies using both obsidian and pottery. The use of multiple material types is advantageous in this research as they highlight exchange patterns of differing intensities at different spatial scales.…”

Digging the topology of rock art in northwestern Patagonia

“…The use of analytical methods derived from network science has undeniably hatched over the last decade, bringing together several disciplines as mathematics, physics, computer science, biology and social sciences. Interdisciplinary approaches to the study of past and present societies, from the point of view of complex systems, have started to dominate the tendency towards the creation of knowledge in anthropology [1,2,3] and archaeology [4,5,6,7,8]. Within this context, the application of analytical methods derived from graph theory and from social network analysis to the study of archaeological problems has been increasing at a fast step.…”