Construction of a Support Environment for Drafting and Translating Statutory Sentences: Utilization of a Terminology and a Translation Memory

本研究的目的是建立技术和环境的汇编和使用的术语和翻译记忆，以支持起草和翻译日本法定句子。我们编制了一个文本语料库，包括二战后颁布的所有日本法案（10，067项法案），一个双语语料库，包括美国占领期间颁布的1，624项法案（156，562个双语句子），以及一个翻译记忆库，包括259项法规，即：法案、内阁命令和条例（147 119句双语）。此外，我们已经开发出的方法提取法律的条款，通过使用浅层分析（组块）或文档频率，从双语术语的语义类别，和法定术语的定义，以及它们之间的语义关系。

The purpose of this research is to establish technologies and environments for compilation and utilization of a terminology and a translation memory to support drafting and translating Japanese statutory sentences. We have complied a text corpus including all Japanese acts (10,067 acts) enacted after the World War II, a bilingual corpus including 1,624 acts (156,562 bilingual sentences) enacted during the period of U.S. occupation, and a translation memory including 259 statutes, i.e., acts, cabinet orders and regulations (147,119 bilingual sentences). In addition, we have developed methods for extracting legal terms by using shallow parsing (chunking) or document frequency, semantic category from bilingual terms, and definitions of statutory terms as well as semantic relations between them.