Agent-based computational models (ABMs) offer a distinctive approach to social science, characterized by their generative nature. Unlike inductive or deductive methods, ABMs simulate decentralized interactions among heterogeneous agents to generate macroscopic regularities. These models allow for the study of complex social phenomena, such as wealth distribution, price equilibria, and cultural patterns, by simulating bottom-up processes. ABMs are particularly useful for addressing interdisciplinary questions and testing the robustness of theories under varying conditions. They enable the examination of how individual behaviors lead to macroscopic outcomes and can be used to evaluate competing models through empirical data. ABMs also challenge traditional notions of rationality and equilibrium, decoupling individual rationality from macroscopic outcomes. Furthermore, they provide a framework for interpreting society as a distributed computational system, raising foundational issues related to intractability and undecidability. ABMs are also valuable for empirical research, as demonstrated by projects like the Artificial Anasazi, which simulate historical demographic patterns. They allow for the testing of hypotheses about social processes and can generate stylized facts about social phenomena. ABMs also contribute to connectionist social science by modeling decentralized, dynamic systems. They offer a new methodology for interdisciplinary research, enabling the study of complex social systems that involve multiple interacting factors. ABMs can also be used to stress-test dominant theories, revealing their limitations. They provide a powerful tool for understanding how individual behaviors lead to macroscopic outcomes and can be used to explore the effectiveness of different social conventions. Overall, ABMs represent a significant advancement in social science, offering a new way to study complex systems and generate insights into social dynamics.Agent-based computational models (ABMs) offer a distinctive approach to social science, characterized by their generative nature. Unlike inductive or deductive methods, ABMs simulate decentralized interactions among heterogeneous agents to generate macroscopic regularities. These models allow for the study of complex social phenomena, such as wealth distribution, price equilibria, and cultural patterns, by simulating bottom-up processes. ABMs are particularly useful for addressing interdisciplinary questions and testing the robustness of theories under varying conditions. They enable the examination of how individual behaviors lead to macroscopic outcomes and can be used to evaluate competing models through empirical data. ABMs also challenge traditional notions of rationality and equilibrium, decoupling individual rationality from macroscopic outcomes. Furthermore, they provide a framework for interpreting society as a distributed computational system, raising foundational issues related to intractability and undecidability. ABMs are also valuable for empirical research, as demonstrated by projects like the Artificial Anasazi, which simulate historical demographic patterns. They allow for the testing of hypotheses about social processes and can generate stylized facts about social phenomena. ABMs also contribute to connectionist social science by modeling decentralized, dynamic systems. They offer a new methodology for interdisciplinary research, enabling the study of complex social systems that involve multiple interacting factors. ABMs can also be used to stress-test dominant theories, revealing their limitations. They provide a powerful tool for understanding how individual behaviors lead to macroscopic outcomes and can be used to explore the effectiveness of different social conventions. Overall, ABMs represent a significant advancement in social science, offering a new way to study complex systems and generate insights into social dynamics.