What Sociologists Should Know About Complexity

Complexity science offers a new lens for sociologists to examine social systems. While traditional sociology often focuses on social structures, institutions, and human interactions, complexity science provides a formal, model-based approach to understanding how these micro-level interactions aggregate into macro-level phenomena. The author argues that complexity science is not a replacement for existing sociological methods but rather a valuable addition to the researcher's toolkit, enabling the study of unasked questions and complex, non-stationary social dynamics.

Complex systems are characterized by four key attributes: diversity among entities, networked interactions, interdependent behaviors, and the capacity for adaptation or learning. These systems are often described as being "between ordered and random" (BOAR) and "difficult to explain, evolve, engineer, or predict" (DEEEP). The author stresses that complexity is a phenomenological property; a system may produce simple patterns in some contexts and complex, unpredictable behaviors in others. Consequently, there is no single measure of complexity, but rather a collection of metrics—such as Kolmogorov complexity or excess entropy—that capture different facets of system behavior.

Central to this approach is the "aggregation diagram," which shifts the focus from variables to agents. In this framework, individual agents with specific behaviors are situated in a social space and interact over time. These micro-level actions aggregate to produce macro-level outcomes, which in turn can influence future micro-level behaviors. This perspective aligns with sociological interests in emergence and self-organization, where macro-level properties (such as social norms or market trends) exist independently of the individual members and may exhibit their own distinct dynamics.