Computational Modeling of the Social Complexity in Göbekli Tepe and Earliest Neolithic Communities

Related publication: Bagheri-Jebelli, N., Kennedy, W.G., “Social Network Analysis of Gobekli Tepe: Belief Propagation in 10th Millennium BC”, International Conference on Social Computing, Behavioral-Cultural Modeling, and Prediction and Behavior Representation in Modeling and Simulation, Washington, D.C., USA, 2020.

Goal: For years, we believed that the transformation of human communities from simple nomadic hunter-gatherer societies to complex sedentary societies occurred due to the expansion of cultivation and the rise of agriculture. Accepting this perspective as premise to years of archaeological research limited the field in investigating alternative explanations to the significant socio-cultural transition which potentially led to years of collaborative efforts in the Mesopotamian region of West Asia. Göbekli Tepe, a pre-pottery Neolithic site in Turkey, requires such alternative explanations for the Neolithic Revolution. Since the site was discovered, various excavations and analyses have been conducted. However, all such investigations include traditional archaeological methods and expectations. This research proposes a multidisciplinary theory of the development of the social complexity of Göbekli Tepe and the contemporary communities in the Urfa region prior to the rise of agriculture. Understanding the interplay of factors resulting in group identity and social cohesion shed light on the reasons that motivated the collective action of hunter-gatherer societies of the region to construct this communal space for ritualistic purposes. This research, rooted in complex system theory, pioneers a novel exploration of the nonlinear dynamics of humanhuman and human-environment interactions among Neolithic communities. Utilizing a multi-method computational approach, it delves into group identity and social cohesion, departing from prior studies confined to single quantitative methods. The study employs quantitative and qualitative socio-cultural and environmental data to construct an artificial society, unveiling nuanced factors shaping interactions. Through agent-based modeling and social network analysis, it sheds light on the dynamics of trust, cooperation, leadership emergence, belief propagation, and universal narratives, offering insights into historical and contemporary societal behaviors.

The model underwent continuous testing during implementation to ensure replicability of observed data, with a conceptualization involving three phases: agent-environmental dynamics, agent behavior, and analysis and interpretation of outcomes. Phase I focuses on agent dynamics during hunting and gathering, addressing survival, prosperity, and subsistence patterns, utilizing diverse data sources for model design. Phase II introduces sociopsychological elements like trust and cooperation, enhancing complexity and discovery pace. In phase III, results are analyzed to reveal patterns and insights on Göbekli Tepe’s dynamics, including group identity, social cohesion, and belief propagation. Social network analysis quantifies structural cohesion, identifies leaders, explores group identity factors, and compares parameters affecting subsistence patterns, combining socio-cultural factors influencing decision-making and adaptive behavior.

Justified results include (i) demonstration of a consistent shift from hunting to gathering in simulated societies, showcasing adaptability and resilience, with insights into potential agricultural emergence from extended wild grain gatherings; (ii) improved understanding of the essential role of trust and cooperation in community formation among hunter-gatherers, impacting network connectivity and emphasizing their influence on the construction of Göbekli Tepe and early human societies; (iii) novel findings that suggest belief propagation in Neolithic communities, indicating a shared belief system’s crucial role in fostering collective identity and influencing the emergence of leaders, contributing to monumental construction at Göbekli Tepe; and, (iv) exploration of the development of universal narratives, demonstrating that trust thresholds influence narrative attachment growth, providing insights into symbolic motivations behind hunter-gatherer monumental structures and emphasizing storytelling’s pivotal role in fostering extensive cooperation.

This dissertation’s research, grounded in complex system theory, innovatively integrates computational methods to explore group identity and social cohesion among Neolithic communities, surpassing prior studies relying on single quantitative methods. The consistent patterns across experiments demonstrate the adaptive resilience of simulated societies transitioning to gathering. Trust emerges as a crucial factor influencing cooperation, social cohesion, and group identity, with implications for modern societal dynamics and policymaking. The study also delves into the emergence of leaders and belief propagation dynamics, offering insights relevant to historical and contemporary societal behaviors. This interdisciplinary computational approach advances scientific understanding, with potential applications in anthropology, archaeology, and policy-making. Collaboration with the Göbekli Tepe archaeological team enhances study validity, fostering international research partnerships and future discoveries.

Conclusion: This study built upon prior agent-based models in archaeology while pioneering the exploration of group identity and social cohesion within the specific context of Göbekli Tepe and the communities in the Urfa region, utilizing social network analysis. It demonstrated the dynamics of human-human and human-environment interactions in a complex adaptive system, where interconnected components interacted at different spatial and temporal scales. As a methodological step in implementing theories of behavior, the model was tested throughout its implementation to ensure the replication of observed data. The findings of the first experiment exploring the hunting versus gathering, showed a consistent pattern emerging as gathering became the primary food source, signaling the adaptability of simulated societies. The model effectively captured the gradual shift from hunting to gathering, showcasing communities’ resilience and learning. The model also demonstrated the adaptive nature of hunter-gatherer communities, incorporating grains into their diets over time, offering insights into the potential emergence of agriculture as a byproduct of extended gatherings focused on wild grains. The results of the second experiment on building trust and cooperation provided a deeper understanding of their pivotal role in community formation which is essential for interpreting observed network dynamics. The findings show that trust restrictions impact network connectivity and cooperation levels, highlighting the conditions conducive to collaborative efforts among hunter-gatherers. These insights shed light on the environmental factors influencing the construction of Göbekli Tepe, emphasizing the interplay between trust, cooperation, and community structures in early human societies.

The findings of the third experiment presented the possible propagation of belief through the Neolithic communities of the Urfa region. The presence of a shared belief system at Göbekli Tepe could have played a crucial role in belief propagation, facilitating the spread of ideas, values, and cultural practices across diverse groups and fostering a collective identity. Within this dynamic, the emergence of leaders may be linked to individuals who played pivotal roles in shaping and disseminating these shared beliefs. Their influence likely contributed to the coordination and organization necessary for the construction of monumental sites like Göbekli Tepe. This highlights the intricate interplay between belief systems, cultural transmission, and the emergence of leadership in the development of ancient hunter-gatherer societies. The results of the fourth experiment demonstrated the development of universal narratives. Exploring narrative evolution in tribes and cultures reveals that an increased trust threshold constrains narrative attachment growth, while reducing it below default settings facilitates significant increases. This deep dive enhances our understanding of simulated society dynamics and sheds light on symbolic motivations driving hunter-gatherers to build monumental structures like Göbekli Tepe, aligning with existing literature on the site’s iconography. The study emphasizes the pivotal role of storytelling in transitioning to a new lifestyle and disseminating religious ideas, indicating its crucial role in fostering extensive cooperation for monumental projects like Göbekli Tepe.

The implications of the research conducted in this dissertation include short-term and long-term interdesciplinary contributions. This research, rooted in complex system theory, pioneers the integration of various computational methods to explore group identity and social cohesion among Neolithic communities, surpassing previous studies that relied on single quantitative methods for specific geographical locations. The findings reveal a consistent pattern across experiments, showcasing the adaptive resilience of simulated societies as they transition towards gathering as the primary food source. Trust emerges as a crucial factor, influencing cooperation, social cohesion, and group identity within societies, with implications for modern societal dynamics and policy-making. Additionally, the study delves into the emergence of leaders and the dynamics of belief propagation, offering insights applicable to both historical and contemporary societal behaviors. The interdisciplinary computational approach contributes to the advancement of scientific understanding and holds potential for broader applications in anthropology, archaeology, and policy-making. Collaboration with the archaeological team at Göbekli Tepe enhances the study’s validity, fostering international research partnerships and future discoveries. While the aforementioned implications are notable, there are limitations to the research. The development of the agent-based model for exploring an archaeological site and ancient society faces limitations due to challenges in data availability and the slow pace of excavation processes, with fragmented and incomplete archaeological records introducing uncertainty into the modeling process. Additionally, the intangible and impermanent nature of concepts like trust, cooperation, belief propagation, and universal narratives poses challenges in validating these elements even with ample data, relying heavily on indirect evidence and interpretations subject to biases. Furthermore, incorporating social psychology theories into agent-based modeling is hindered by inadequate interdisciplinary validation methods, highlighting the need for collaborative efforts to bridge the gap between social psychology and computational modeling for a more comprehensive understanding of socio-psychological behaviors. Despite the limitations, furthering the research is a valuable endeavor. The Göbekli Tepe model’s complexity opens avenues for further exploration, including the examination of model parameters, extension of the environment to include topological and climate factors influencing site abandonment, and analysis to improve understanding of agent networks, with the intention to enhance decisionmaking through reinforcement learning algorithms. Social network analysis concepts like modularity and assortativity provide valuable insights into the socio-cultural dynamics at Göbekli Tepe, revealing potential divisions or roles within the community and shedding light on social ties and interactions among cultural or functional groups. Multi-agent reinforcement learning algorithms, addressing cooperative, competitive, and mixed tasks, offer a promising approach to studying emergent behaviors, complexity in agent communication, and enhancing agent decision-making through interactions with the environment and other agents. The exploration of various reinforcement learning algorithms aims to identify the most suitable method for the Göbekli Tepe case.