The Modeling of Cognitive Relevance and the Actural Study based on the Experiments of Sentence Comprehension

逻辑推理是人类智能的基础，但认知主体的逻辑处理系统不同于命题逻辑，特别是在解决与对位有关的问题时。这在一定程度上是因为，由于我们的能力有限，认知代理无法获得所有必要的推理信息。因此，我们的认知系统被期望找到狭窄的最优关联信息。关联理论（Sperber and Wilson 1986）建立在关联的定义和两个关联原则的基础上。其中一个原则被称为认知原则，这意味着人类的认知是与相关性最大化相适应的。本研究提出了基于认知原理的多值逻辑框架，并揭示了信息的确定性和相关性对日常推理的影响，包括对日语各种类型条件句的解释过程的影响。

Logical reasoning is fundamental to human intelligence, but the logical processing system of cognitive agents is different from propositional logic, especially on solving problems which are linked with the counterposition. This is partly because a cognitive agent cannot acquire every essential information for reasoning because of our limited capacity. Therefore, our cognitive system is expected to find the narrow optimal relevance information. Relevance theory (Sperber and Wilson 1986) is based on a definition of relevance and two principles of relevance. One of the principles, called Cognitive Principle, means that human cognition is geared to the maximisation of relevance. This research proposes the frame work of multiple-valued logic based on the philosophy of the cognitive principle, and reveals the effect of both settledness and relevance of information on everyday reasoning including of the interpretation process of the various types of conditionals in Japanese.