Making Sense of Analogies in Metacat

This paper outlines the main ideas and objectives of the Metacat project, an extension of the Copycat computer model of analogy-making and high-level perception. The principal features of Metacat that allow it to make sense of analogies suggested to it by the user are described using a simple example. Introduction The Copycat computer model of analogymaking and high-level perception was originally developed by Hofstadter & Mitchell as a computational model of subcognitive mechanisms underlying human cognition, in which the notion of fluid concepts plays a central role. Copycat models the process of analogy-making within a strippeddown microworld of tiny, idealized situations represented as short strings of letters. For example, a typical Copycat problem is the following: "If abc changes to abd, how does mrrjjj change in an analogous way?" This microworld, though austere, harbors a surprisingly rich variety of subtle problems in which a wide range of answers is almost always possible-..