Worlds of change : counterfactual reasoning and causation / Charles L. Ortiz.

Format:

Book

Manuscript

Microformat

Thesis/Dissertation

Author/Creator:

Ortiz, Charles L.

Contributor:

University of Pennsylvania.

Language:

English

Subjects (All):

Penn dissertations--Computer and information science.

Computer and information science--Penn dissertations.

Local Subjects:

Penn dissertations--Computer and information science.

Computer and information science--Penn dissertations.

Physical Description:

xiv, 294 pages : illustrations ; 29 cm

Production:

1996.

Summary:

The development of a commonsense theory of causation has often been pursued along a number of, somewhat orthogonal, directions. The work contained herein examines the role that counterfactual reasoning can play within such a theory. The sorts of inferences one might draw in the course of causal attribution is first examined; this pre-theoretic analysis is conducted within the context of a rich microworld involving a number of agents engaged in purposeful behavior. A new theory of change, Explanatory Update Theory (EUT) is developed in which: (1) a syntactic version of belief updating underlies the semantics of counterfactuals and in which events and time are explicitly represented at the object level; (2) a new solution to the frame problem is presented based on a notion of support for a proposition which corrects a number of problems with existing approaches involving unexplainable events, the need for persistence rules, explanation in the context of incompletely specified causal chains, and event ramifications; (3) the solution to the frame problem is integrated with the semantics for counterfactuals; and (4) the epistemic preferences that underly the choice of alternative worlds in accommodating a counterfactual supposition in the course of causal attribution is carefully articulated. EUT is then shown to correctly handle a number of standard benchmark problems as well as a number of new benchmarks emerging from the microworld study. A commonsense causal language is formalized consisting of terms such as preventing, enabling, maintaining, letting, helping, hindering, and so forth. The utility of counterfactual reasoning in supporting the production of causal explanations involving negative event descriptions is also demonstrated. A stratified view of causal reasoning is shown to emerge in which the utility of counterfactuals manifests itself in two ways. First, as useful tools in isolating the role that an event played in some nexus of events, and also as providing an efficient means for combining both causal and non-causal knowledge in the production of causal explanations.

Notes:

Thesis (Ph.D. in Computer and Information Science) -- University of Pennsylvania, 1996.

Includes bibliographical references and index.

Local Notes:

University Microfilms order no.: 97-12985.

OCLC:

187469417