ARTIFICIAL INTELLIGENCE 259 Analog Retrieval by Constraint Satisfaction Paul Thagard Cognitive Science Laboratory, Princeton University, 221 Nassau St., Princeton, NJ 08542, USA Keith J. Holyoak Psychology Department, University of California, Los Angeles, CA 90024, USA Greg Nelson Cognitive Science Laboratory, Princeton University, 221 Nassau St., Princeton, NJ 08542, USA David Gochfeld Oberlin College, Oberlin, OH 44074, USA ABSTRACT We describe a computational model of how analogs are retrieved from memory using simultaneous satisfaction of a set of semantic, structural, and pragmatic constraints. The model is based on psychological evidence suggesting that human memory retrieval tends to favor analogs that have several kinds of correspondences with the structure that prompts retrieval: semantic similarity, isomorphism, and pragmatic relevance. We describe ARCS, a program that demonstrates how these constraints can be used to select relevant analogs by forming a network of hypotheses and attempting to satisfy the constraints simultaneously. ARCS has been tested on several data bases that display both its psychological plausibility and computational power. 1. Introduction Analogy is ubiquitous in human thinking, in diverse areas that range from practical problem solving to scientific explanation to literary embellishment. This paper addresses one of the central questions related to understanding analogy: How can relevant analogs be retrieved from memory for potential use? Numerous computational models have been proposed for analog retrieval (sometimes also called access or recognition) using mechanisms such as similari- Artificial Intelligence 46 (1990) 259-310 0004-3702/90/$03.50 © 1990 -- Elsevier Science Publishers B.V. (North-Holland)