ABSTRACT
The Western natural sciences have a long methodological tradition based on bivalent reasoning and mathematical formulation. The former paradigm adopts an either-or approach to the examination of the phenomena and the latter often assumes a full correspondence between the mathematical and the real world. Despite their fairly successful results in the natural sciences, these traditions have aroused problems when applied to human reasoning or intelligent computer models. The bivalent models are too coarse for representing complex phenomena, and the conventional mathematical models have often been too complicated in any discipline. In particular, these methods have been more or less controversial when applied in the human sciences (social sciences, behavioral sciences, economics, etc.).
