Implementation of the abstractional concept mapping theory for the evaluation of four-part geometric analogies

In this paper we discuss work in progress on the application of an existing model of analogical reason ing, referred to as the Abstractional Concept Mapping Theory, to four-part geometric analogy problems. Work until the present has involved development and characterization of a formal model, as yet untested by implementation. We have chosen in our work to have the first application of the model be four-part geometric analogy problems for the following reasons: (1) it provides a relatively "simple" domain of knowledge to be encoded, (2) it provides the opportunity to compare results with an earlier, well-known program for solving four-part geometric analogies, and (3) it has potential practical application in intelligence testing involving the use of geometric analogies.