New graphical Frameworks for Causal Discovery and Control

Twenty years ago Pearl suggested using the framework of Bayesian Networks to formally define collections of causal hypotheses in terms of hypotheses about the effects of applying a control. The same semantics had also been proposed by Spirtes Glymour and Sheins to guide the search for putative causal hypotheses between high dimensional multivariate observational data. Although this work was seminal in its time I will argue here that the proposed semantics certainly do not underpin a general definition of causality and are not entiely applicable to many scientific domains. In particular both its suppression of the most obvious of axioms – that a cause happens before an effect and its requirement that a “cause” be associated with a random variable are each problematic. I will propose various alternative expressions of causal hypotheses more expressive of the actual causal hypotheses than their BN analogues when used in certain contexts: especially dynamic ones. I will give illustrations from applications associated with fMRI imaging, the analysis of gene expression data & longitudinal behavioral studies.