ABSTRACT
The binary contrasts up to this point have provided evidence for the integration of independent and continuous sources of information in bimodal speech perception. The issue addressed in the current chapter concerns the nature of the integration process. Without a doubt, this question is the most difficult to answer primarily because of its intricate dependence on quantitative models and tests (Falmagne, 1985). Background on the nature of testing integration rules is provided because the question is a relatively novel one, not only for speech perception, but pattern identification in any domain. Our goal is to provide a reasonable foundation for testing various integration rules given multiple sources of continuous information. We consider three classes of integration rules: additive, minimization, and multiplication. The first class has been most successfully applied in the area of person impression. The latter two classes are based on theoretical work in the field of fuzzy sets.
