A methodology is presented for qualitative modeling and simulation of the temporal behavior of individual variables and the composite process. The temporal behavior of a variable in a rt windowed time scale is captured categorically using principal-component analysis, while the process is modeled as a causal diagraph with intel acting and recycle nodes. Rigorous, rather than nd hoc, reasoning rules can be devised for the causal diagraph using a learning mechanism. Quantitative information can also be incorporated into the method with the introduction of fuzzy c-means clustering without compromising the cognitive level of information embedded in the nodes and the digraph.