生成语法的目标与体系——与人工智能途径的对比

A Comparison of the TG and AI Approaches to Language

本文对生成语法和人工智能这两种语言研究范式进行系统的对比分析,涉及理论目标、理论模型、语料来源、哲学归属、结构分析、词与词库等6个方面。基本结论是:两者的主要分歧源于各自的理论目标,体现在理论体系的各个方面。两者互不适应,分道扬镳在所难免。不必采用扬此抑彼的态度,也无需试图对两者进行深度的调和与融合。两种研究各有不可替代的价值,在不同领域,以不同方式,或深化我们对语言本体机理的认识,或促进语言信息技术的应用。二者间不存在范式更替问题,共同发展不仅可能,而且必要。

This paper compares transformational-generative grammar and artificial intelligence, two prominent approaches in contemporary linguistic research. The comparison spans over six major aspects, including their objectives, theoretical models, data sources, philosophical underpinnings,structural analyses, and the role of the lexicon. Based on the comparison, it is concluded that the two approaches are fundamentally different, developing in different directions. However, the two approaches do not stand in a position for a paradigm shift. Each has its own contribution to make,with transformational-generative grammar in the realm of the inner workings of the language, and AI in the area of language-related information technologies.