A Basic Study of Legal Expert System--Artificial Intelligence Aproach to Legal Reasoning--

本研究的目的是通过对实证法语言的分析，厘清法律知识的逻辑结构，并为能够根据其实践进行法律正当性推理的法律专家系统奠定基础。为了阐明和论证这些原理和方法，本研究的目的是在一个由知识库、推理机和用户界面组成的人工智能工作站上制作一个实验原型。我们几乎完全完成了这一研究目标。(1)在法律知识结构的澄清方面，我们从法律要求和法律效力的角度对《联合国国际货物销售合同公约》(部分)的条款进行了分析，明确了法律规则之间的逻辑联系。(2)在人工智能工作站PSI-II上，我们建立了由上述领域的法律规则和元法律规则的复合谓词逻辑公式组成的法律知识库。(3)就推理机而言，我们取得了以下成果：(A)我们开发了一个推理机，它执行推理以确定适用的法律规则，同时区别对待这些推理；(B)我们已经建立了一个系统来解释所执行的影响过程；(C)我们已经使用法律同义词词典结构实施了一个引证推理的工程模型，并审查了其有效性和相关问题。此外，我们还阐明了法律电子概念词典的某些原则和方法，可以说，我们已经成功地考察了该领域的逻辑、法哲学、法律社会学、民法、信息工程和知识工程基础。它使我们相信，以这项工作为基础，我们可以进行研究，以开发一个全面的法律专家系统。

The object of this research is to clarify the logical structure of legal knowledge through the analysis of the language of positive law, and to develop a basis for a legal expert system which can carry out inferencing for legal justification according to its praxis. In order to clarify and demonstrate these principles and methods, this research aims at making an experimental prototype on an AI workstation which consists of a knowledge base, an inference engine and a user interface. We have almost completely accomplished this research objective.(1) As far as the clarification of the structure of the legal knowledge is concerned, we have analyzed the articles of the United Nations Convention on Contracts for the International Sale of Goods (part) in terms of the legal requirements and legal effects and made clear logical connections among the legal rules.(2)On the AI-workstation PSI-II, we have made a legal-knowledge base composed of compound predicate logical formulae of legal rules and meta-legal rules in the above areas.(3) As fat as the inference engine is concerned, we have achieved the following : (a) we have developed an inference engine which performs inferencing to determine the applicable legal rules while treating those inferences dedctively ; (b) we have built a system to explain the inferencial processes performed ; (c) we have implemented an engineering model of adductive reasoning using legal thesaurus structures and have examined its efficacy and relevant problems. Further more we have clarified certain principles and methods of the electronic concept dictionary of low.We can say that we have successfully examined the logical, legal-philosophical, legal-sociological, civil law, information-engineering and knowledge-engineering foundations of the field. The has convinced us that using this work as a basis we can undertake research toward the development of a full-scale legal expert system.