Abstract

Using Bayesian network causal models, we provide a simple general account of probabilistic causal interaction. We also detail problems in the leading accounts by Ellery Eells, and any others which require valence reversals, contextual unanimity, or average effects.