Bayesian Nets Are All There Is To Causal Dependence

There are too many theories of causation to get into the focus of a small paper. But there are two in which I have a natural interest since they look almost the same: namely the theory of Clark Glymour, Peter Spirtes, and Richard Scheines, so vigorously developed since 19831 and most richly stated in Spirtes et al. (1993) (whence I shall refer to it as the SGS theory), and my own theory, published since 1978 in a somewhat irregular way. They look almost the same, but the underlying conceptions turn out to be quite dissimilar. Hence, the original idea for this paper was a modest one: simply to compare the philosophical basics of the two theories. However, no paper without a thesis! Therefore I have sharpened my comparison to the thesis written right into the title. The plan of the paper is simple. Section 2 sets out the formal theory of Bayesian nets in an almost informal way, and section 3 analyses the philosophical differences hidden in the common grounds. Section 4 briefly extends the comparison to the treatment of actions or interventions.