Steps Toward Automatic Theory Formation

This paper describes a theory formation system which can discover a p a r t i a l axiomization of a data base represented as extensional ly defined binary r e l a t i o n s . The system f i r s t discovers a l l possible in tens ional d e f i n i t i o n s of each binary re la t i on in terms of the others. It then determines a minimal set of these re la t ions from which the others can be def ined. It then attempts to discover a l l the ways the re la t i ons of t h i s minimal set can in teract wi th each other, thus generating a set of inference ru les . Although the system was o r i g i n a l l y designed to explore automatic techniques for theory construct ion for question-answering systems, i t is current ly being expanded to funct ion as a symbiotic system to help soc ia l s c i en t i s t s explore cer ta in kinds of data bases.