Chater and Christiansen (CC) offer a compelling account of how socially propagated ad hoc solutions to local problems involving social coordination can promote increased cultural complexity. However, their account does not explain why such transitions often appear abruptly rather than gradually and unevenly across different regions, nor does it explicitly allow for possible reverses in complexity. We suggest that a more complete account of the dynamics of cultural change is provided by complex dynamical systems and the application of this approach to social processes (Nowak & Vallacher, Reference Nowak and Vallacher1998; Vallacher & Nowak, Reference Vallacher and Nowak1997).

Within the complex dynamical systems framework, rapid social change does not represent linear progression but rather follows the scenario of phase transitions in physics. In this scenario, change occurs in a nonlinear fashion, with growing clusters of “new” forming and growing in the sea of “old.” The rapid transition from water to steam, for example, occurs as bubbles of the new (steam) form and grow within the old (water), with the two states coexisting in space. Analogously, agent-based modeling has shown that new trends in a society (e.g., fashions, opinions, and norms) do not happen in a linear progression with everyone progressively adopting the new trend but rather develop within small clusters of interconnected individuals (e.g., Nowak, Szamrej, & Latané, Reference Nowak, Szamrej and Latané1990). A social transition occurs as these minority clusters grow and spread to neighboring individuals in social space, with the trend becoming the majority. In social systems, however, a complete transition is rare; the system instead stabilizes in a semi-stable state – an attractor – with the old continuing to exist in clusters of strong, interconnected individuals who resist the new. During social transitions, then, the “old” and the “new” coexist as clusters separated in social space. If conditions subsequently favor the old, the clusters of old can serve as seeds for reversing the transition, with the old becoming the majority.

Empirical data concerning the socioeconomic transition in Poland fits this scenario (Nowak & Vallacher, Reference Nowak and Vallacher2019). After the collapse of central-government control in the 1980s, clusters formed of individuals who supported pro-market reforms and pro-democracy initiatives. These separate innovations, both predicted by the average educational level at the county level, became integrated, allowing the clusters to merge and promote a higher level of cultural complexity. Although an election favored the new innovations, the old preferences survived in clusters of strong and highly interconnected individuals who could resist the influence of the new. Because conditions subsequently undermined preference for the new (e.g., increased unemployment), the next election favored the old. This reversal was faster than the original transition because the spatial structure of the old was already in place. Oscillation between preference for the old and the new was observed for several election cycles, representing switches between alternative attractors.

We suggest that the account epitomized in the Polish case captures rapid social change in general, including the innovations to solve problems of social coordination that are the focus of CC’s account. The progressive stages they identify represent semi-stable states that function as attractors. In effect, this process reflects a scenario of punctuated equilibrium, with periods of stability separated by periods of rapid change to new attractors. The attractors at each stage are unlikely to represent a uniform level of cultural complexity but rather a mix of clusters with different levels of complexity.

Because innovations originate locally (e.g., within clusters), different innovations may remain independent or be mutually incompatible. For example, the development of artificial intelligence, which requires tremendous energy consumption, may be at odds with the development of new technologies that emphasize energy conservation. Similarly, the development of electric vehicles is at odds with developing mass public transportation that obviates the need for personal vehicles. If different innovations can be integrated in terms of a higher-order attractor – as was observed in the Poland experience – progress toward greater complexity can occur. But if innovations cannot be integrated with respect to a higher stage of complexity, the system will remain stabilized at lower levels of complexity, with the innovations seen as distinct or incompatible rather than having a common goal or purpose. This is analogous to how a person’s failure to integrate their specific actions (e.g., motor movements) in terms of a superordinate goal or purpose keeps personal understanding and control at the level of the person’s specific actions (e.g., Carver & Scheier, Reference Carver and Scheier1998; Vallacher & Wegner, Reference Vallacher and Wegner1987). Progress in cultural complexity thus is not a given but rather a fragile achievement that is open to reversal to a lower level of complexity. For increases in cultural complexity to be maintained, the clusters of interconnected individuals must find ways to grow and achieve integration with other clusters in the society.

An intriguing issue raised by CC’s account is the relationship between stages of cultural evolution and the stability of their respective attractors. On one hand, there may be an inverse relationship between increased complexity and stability. Thus, while early stages are based on simple processes that are inherent in human nature (e.g., learning by observation and language), later stages associated with greater complexity put increasing demands on human institutions for their maintenance (e.g., government administration and advanced technologies). These institutions are more sensitive to disruption, creating the potential for regression to earlier stages of cultural complexity. But on the other hand, the greater redundancy and safeguards associated with attractors provided by institutions may enable the complexity associated with later stages to be preserved. The relative merits of these competing perspectives represent a fertile ground for research.

In sum, CC’s account of stages in the progression toward cultural complexity could be made more precise by specifying the dynamics responsible for this progression. Such refinement would also allow for predictions regarding factors that determine whether, and to what extent, there is unevenness across different segments of society in progressive cultural complexity and whether reversals of complexity are likely. The tools associated with complex dynamical systems (e.g., agent-based modeling and attractor neural networks) provide proven means for investigating these dynamic processes and their manifestation for diverse social processes.