Surviving in the early Neolithic. Causal networks and complex systems in archaeology

• Bayesian networks model complex survival factors in early farming communities. • Two complementary models: theory-driven and data-driven. • Environment shaped herding and hunting, while social structure influences agriculture. • Non-linear relationships reveal how ecology and society jointly affect survival. • Probabilistic approach handles uncertainty in archaeological data interpretation. Difficulties surrounding the reconstruction of social systems in past communities have propitiated the development of multiple social theories and a variety of approaches to explain archaeological remains. The Bayesian Network approach has proved to be a crucial tool to model uncertainty and probability to estimate parameters and predict the effects of social decisions, even when some data entries are missing. This paper has the principal objective to present a research study centered on exploring how prehistoric early farmers survived in their environmental context by suggesting a causal complex model of a socio-ecological system. To achieve this, two different causal models are proposed, both based on probabilistic Bayesian Networks, one built from expert knowledge and the other learned from ethnoarchaeological data. These models are used to define what variables would have been relevant to the socioeconomic organization of early Neolithic communities and to predict their behavior and social decisions in hypothetical case scenarios. The ultimate outcome is exploring the use of the Bayesian Network for investigating socio-ecological systems and defining its potentialities as a research method.