|
Abstract : |
This paper is concerned with the problem of determining the indirect effects or ramifications of actions. We argue that the standard framework in which background knowledge is given in the form of state constraints is inadequate and that background knowledge should instead be given in the form of "causal laws. " We represent "causal laws " first as inference rules and later as sentences in a modal, conditional logic C flat. For the framework with "causal laws, " we propose a simple fixpoint condition defining the possible next states after performing an action. This fixpoint condition guarantees minimal change between states, but also enforces the requirement that changes be "caused. " Ramification and qualification constraints can be expressed as "causal laws." 1, |
