By J. Daniel Rogers and Wendy H. Cegielski
Last week the Proceedings of the National Academy of Science published our commentary, “Building a better past with the help of agent-based modeling” (www.pnas.org/cgi/doi/10.1073/pnas.1718277114). Our topic is computational modeling and trends in how archaeology and other historical sciences conduct research. We focus on a methodology called agent-based modeling. This is a kind of complex simulation that creates agents, typically having characteristics of humans. The agents interact, make decisions, and changes take place over time. The results are accumulated and compared to real-world information about some aspect of a social system, depending on the question being investigated.
Our back story for writing on this topic comes from a convergence of perspectives plus the inevitable power of opportunity. First of all, although we work at different institutions (the Smithsonian and Arizona State University) and study different topics in different parts of the world, we both come from an American tradition of archaeological research that values both scientific and humanistic perspectives. While our graduate training took place 40 years apart, we still share foundational thinking, although our knowledge base is quite different. A good collaboration creates an opportunity to make the most of shared and divergent perspectives and allows an awareness of certain trends that we might not have recognized individually.
In the PNAS Opinion we note that the current generation of sophisticated computational methods creates the context for a major change in thinking, similar to a paradigm shift. We are not the only ones to comment on the significance of simulations and ABMs (Banks 2002; Epstein 2006; Farmer and Foley 2009; Lake 2015). Until recently, however, the conceptual revolution has remained marginal, largely due to the scarcity of practitioners. Now there are several universities training interdisciplinary scholars who make routine use of high-level computational methods. The emerging potential of computational modeling provides the methodology from which to consider how we do our work as archaeologists.
Our original title was Building a Better Past, which was a bit too grand but it did reflect our interest in engaging with the social sciences and other historical sciences, ranging from paleobiology to astrophysics. To some extent we see archaeology as a proxy for the social sciences more broadly and agent-based modeling as a proxy for a wider variety of computational methods. The Opinion published in PNAS was preceded by an article we wrote on trends in the archaeological use of agent-based modeling (Cegielski and Rogers 2016). In that article we analyzed publishing patterns, topic areas, and gave examples and descriptions of the fundamental workings of an agent-based model. From that survey of the literature it was easy to recognize that as a profession, archaeology had crossed a threshold. Computer simulations were no longer seen as a novelty, but rather as an emerging core methodology.
Bankes, Steven C. 2002 Agent-based Modeling: A Revolution? Proceedings of the U.S. National Academy of Sciences of the United States of America 99, no. Suppl 3: 7199–7200. https://doi.org/10.1073/pnas.072081299.
Cegielski, Wendy H. and J. Daniel Rogers. 2016 Rethinking the role of agent-based modeling in archaeology. Journal of Anthropological Archaeology 41:283-298. DOI: 10.1016/j.jaa.2016.01.009
Farmer, J. Doyne, and Duncan Foley. 2009 The Economy Needs Agent-based Modelling.” Nature 460: 685–86.
Lake, Mark W. 2015 Explaining the Past with ABM: On Modelling Philosophy. In Agent-based Modeling and Simulation in Archaeology, edited by Gabriel Wurzer, Kerstin Kowarik, and Hans Reschreiter, 3–35. London: Springer.
Rogers, J. Daniel and Wendy H. Cegielski. 2017 Building a better past with the help of agent-based modeling. Proceedings of the U.S. National Academy of Science 114 (49): 12841-12844. DOI: 10.1073/pnas.1718277114
A free download is available at: www.pnas.org/cgi/doi/10.1073/pnas.1718277114