Uncovering the Causal Relationships Between Ecology and Social Behaviors in Prehistoric Agropastoral Societies: a Bayesian Network Approach

OH, Man-Suk

This study explores how prehistoric agropastoral societies adapted to environmental constraints for survival, focusing on the causal relationships between ecological factors, subsistence strategies, social organization, and survival decisions. Building on the work of Palacio et al. (2022, Plos one,17(10), p.e0276088), we employ a Bayesian Network approach to uncover the complex interactions among these variables. We use mixed-type data, incorporating both numerical and categorical variables to avoid information loss from data discretization. We apply the causal Bayesian network structure learning algorithm by Spirtes et al. (2000; Causation, Prediction, and Search; MIT Press), which captures both cause-and- effect relationships and associations where causality is uncertain. Finally, by constructing a holistic network that includes all relevant variables, rather than limiting analysis to sub-networks, we provide a comprehensive view of interdependencies and minimize the risk of overlooking potential confounders. The network, learned from 28 variables across 152 trans-historical and cross-cultural small-scale farming societies provided by Palacio et al. (2022), reveals that topological factors—elevation and proximity to the coast—influence climate variables like precipitation and temperature. These climate factors subsequently impact productivity and determine settlement types (e.g., camps, villages, hamlets). Subsistence strategies, including agriculture, hunting, fishing, and gathering, are interrelated, likely because they collectively contribute to a diverse subsistence strategy that enhances survival and resilience. Additionally, topological factors affect hunting and fishing practices. Community size influences agricultural activities and is associated with gathering. Social decision-making factors display specific associations, with reciprocity linked to fishing and temporal migration connected to external exchange. These findings provide new insights into the adaptive strategies of prehistoric societies, highlighting the intricate relationships between environmental and social factors. The Bayesian Network approach offers a robust framework for understanding the complex dynamics that shaped early human communities.

Session 2. Interdisciplinary Approaches to the Study of Ancient Economies ^[info]