“…Studies of causation in artificial intelligence (AI) have primarily focused on statistical (Charniak & Goldman, 1989;Gabrielian & Stickney, 1987;Kim 8c Pearl, 1983;Liu, 1986;Pearl, 1986Pearl, , 1987aPeng & Reggia, 1987;Weiss, Kulikowski, & Amarel, 1978) and logical approaches (Cox & Pietrzykowski, 1986;Fusaoka & Takahashi, 1986;Hagers & Hansson, 1983;Pearl, 1987b;Shoham, 1988). In our view, inductive nature of causation renders these perspectives quite limited as the sole basis for computational models.…”