Induction as a heuristic based on structures of regularities

Abstract Induction is a heuristic method to guess or to discover regularities (constant patterns that things follow such as a large variety of categories or natural laws). However, the main stream of the study of induction focuses not on the heuristic method but on strength evaluation of given inductive arguments. This article describes an attempt to construct a framework of this heuristic method on the basis of the idea that human beings have acquired through evolution a perceptual-cognitive system with the functions: perception of things and certain primitive regularities in the perceived things, extraction of general structures of the perceived regularities, and application of the acquired knowledge of the structures to guess prospective regularities. A way of this application would be the heuristic method. The constructed framework creates inference that satisfies main properties of induction by regarding strength of inductive arguments as rationality of the inference in the framework. Moreover, the framework includes simple cases of analogy, abduction, and induction about causal relationships between events. Some testable predictions to examine this framework are raised, and the results of this framework is compared with those of existing models of induction.