事格语法 (Event-CaseGrammar,简称ECG文法 )以俄语为语料 ,以实现自然语言处理 (naturallan guageprocessing ,NLP)为目的 ,从认知层面切入 ,对俄语句子的语法结构、语义结构以及功能机制进行多层级、多平面的形式化定性描写。我们认为ECG文法有较强的可操作性和可计算性 ,可为NLP提供新的语言学策略和合理的程序保证。本文将围绕三个问题展开分析和研究 :1)ECG文法概述 ;2 )语句知识的形式化表达 ;3)语句信息提取的方法和策略 Event-Case Grammar (ECG Grammar) uses Russian as the corpus and implements naturallan guageprocessing (NLP). From the cognitive level, the grammatical structure, semantic structure and functional mechanism of Russian sentences Multi-level, multi-plane formal qualitative description. We believe that ECG grammar has strong operability and computability, which can provide NLP with new linguistic strategies and reasonable procedural guarantees. This article will focus on three issues analysis and research: 1) ECG grammar; 2) the formal expression of the statement of knowledge; 3) the method and strategy of statement information extraction