Strategies for Dealing with Causal Complexity

Many of the systems studied in contemporary science are causally complex in various ways. Such complexity includes (but is not limited to) cases in which the variables in a system are causally related via a dense network of interacting causes and those exhibiting different behaviors at higher- and lower-levels of description. Causal complexity both renders it difficult in practice to predict the results of interventions, and raises conceptual issues regarding the application of causal concepts to systems that cannot be fully decomposed into modular components. In this symposium, each participant raises a challenge arising from causal complexity and presents a strategy for managing it. The strategies considered include: searching for causal pathways shared by otherwise heterogeneous phenomena, invoking higher-level variables, and abstracting away from causal details. Our discussion moves beyond existing debates regarding the metaphysical respectability of higher- and inter-level causation towards methodological questions about when it is fruitful to model a system at multiple "levels". Through considering real-world examples from a range of scientific domains, we explore how scientific problem solving involves the interplay of considerations related to accuracy and intelligibility and provide a forum for considering the role of causal reasoning in a complex world.