Reference: Iwasaki, Y. Causality and Model Abstraction. Knowledge Systems Laboratory, March, 1989.
Abstract: This paper describes a computational approach, based on the theory of causal ordering, for inferring causality from an acausal, formal description of a phenomena. Causal ordering, first proposed by Simon, in an asymmetric relation among the variables in a self-contained equilibrium or dynamic model, which reflects people's intuitive notion of causal dependency relations among variables. We extended the theory to cover models consisting of a mixture of dynamic and equilibrium equations. When people's intuitive causal understanding of a situation is based on a dynamic description, the causal ordering produced by the extension reflects such intuitive understanding better than that of an equilibrium description. As the number of variables in a system increases and the system becomes more complex, model abstraction becomes essential in reasoning about its behavior. Aggregation of a nearly decomposable dynamic systems is an abstraction technique that provides a formal justification for commonsense abstraction whose application is easily observable in everyday life. The paper examines the close relation between aggregation and causal ordering.
Notes: Revised March 1993.
Full paper available as ps.