Donald Bren School of Information and Computer Sciences

In this paper we describe GLAUBER, an AI system that models the scientific discovery of qualitative empirical laws. We have tested the system on data from the history of early chemistry, and it has rediscovered such concepts as acids, alkalis, and salts, as well as laws relating these concepts. After discussing GLAUBER we examine the program's relation to other discovery systems, particularly methods for conceptual clustering and language acquisition.