试论认知逻辑研究中的若干问题

On some Problems in the Research on Epistemic Logic

认知逻辑研究的目的是为人工智能研究提供有力的工具。然而,目前认知逻辑的研究成果却难于应用于人工智能的实践之中。这种情况在很大程度上是由于认知逻辑自身存在的问题造成的。第一,认知逻辑理论缺少可操作性,人们不能用能行的方法构造认知逻辑系统定理的证明和根据系统所进行的推演。第二,认知逻辑理论所采用的可能世界语义学方法不能对认知语句做出合乎情理的解释。第三,认知逻辑的公理系统并不是人工智能研究所必需的。由于这些问题的存在,逻辑学家需要对原有的方法进行认真的反思。

The purpose of the research on epistemic logic is to provide effective tool for artificial intelligence. However, it is difficult to apply the present theory of epistemic logic to the practice of AI. This embarrassing situation is mainly due to some defects existing in the epistemic logic. Firstly, the epistemic logic lacks of procedural effectivenesss so that people can not construct proofs and deductions effectively according to epistemic logic. Secondly, the semantic method of possible world adopted by epistemic logic can not interpret epistemic sentences properly. Thirdly, the axiomational systems of epistemic logic are not indispensable to the research of AI. Because of the existence of these problems, logicians need to rethink their previous methods seriously.