人工智能研究中的“母语意识”刍议——以对于汉语量词的刻画为案例  

How to Make Computers' Interfaces More Friendly to Native Speakers of Chinese:A Case Study of the Classifiers in Chinese

作　　者：徐英瑾[1] 

机构地区：[1]复旦大学哲学学院,上海200433

出　　处：《社会科学战线》2018年第1期31-40,共10页Social Science Front

基　　金：国家社会科学基金重大项目(15ZDB020);国家社会科学基金一般项目(13BZX023)

摘　　要：从英语思维的角度来讲,现有主流人工智能技术的编程语言与界面语言,对于英语言说者有明显的友好度。隐蔽在人工智能技术的中立面相背后的英语全球霸权,很可能会随着人工智能技术自身的应用得到进一步的扩张。而作为这种趋势的对抗措施,最好的办法就是激活各个非英语民族自身的文化传统,并找到其与合适的人工智能技术刻画方式相结合的方式。汉语中的量词现象,便是一个可以测试上述策略之可行性的具体案例。具体而言,与蒯因基于西语思维对于汉语或日语量词语法功能的刻画方式相对照,一种基于汉语自身韵律学特征的演化历程与认知隐喻理论的量词刻画方案,或许能够更合理地说明量词在汉语或日语中发挥的真实功能,而这种刻画方案自身也可以通过"非公理推演系统"而得到算法化的说明。The expansion of Artificial Intelligence(AI)technology may greatly facilitate the strengthening of the global hegemony of the English language,given that the interfaces and programing languages of mainstream technologies in AI nowadays are mostly designed more friendly to English native speakers.A considerable strategy for balancing this tendency is to build new computing platforms which are supposed to be more friendly to non-Anglophone(especially non-Western)languages.One of the touchstones for testing the feasibility of this strategy is related to the wide currency of classifiers in Chinese or Japanese,which is a linguistic phenomenon that English lacks.With contrast to W.V.Quine and others'attempting to explain classifiers either in terms of numeral components or by virtue of individuating devices,what this paper proposes is a hybrid explanation of the substantial functions of classifiers in terms both of the evolution of the prosodic features of Chinese language and of the metaphor-based features of the Chinese semantic content.Such an explanation could be algorithmized as well,if an appropriate computing tool,such as Wang Pei's Non-axiomatic Reasoning System,is provided.

关 键 词：量词 人工智能 隐喻 韵律学 

分 类 号：B516[哲学宗教—外国哲学]