摘要
哲学家塞尔曾指出,任何一台处理汉语输入的计算机都不能够像真人那样理解汉语,因为计算机无法获取汉语表达式的语义内容。他的这个论证,既预设了语义学和句法学之间的二分,而且也预设了一种外在论式的语义学观点。上述两个预设在本文中都会受到挑战。首先,本文将采纳一种内在论的(并且是准整体论)的语义学进路:根据此进路,意义完全可以在不直接牵涉到和外部对象的关系的情况下而被带入语义网。此外,本文所试图构建的语义学模型也不是句法驱动的——换言之,在该模型中,并没有任何语义中立的原则可以告知系统,复合语义是如何从"原子"语义中生成的。毋宁说,"原子"语义之间的亲和性,便已然在复合语义的构成中扮演了枢纽性角色。需要指出的是,关于语义融合的语义学难题,绝非当代语言学家和语言哲学家的专利。实际上,它早在东汉学者许慎的"六书"论中就已得到触及。根据许论,只要一个汉字的构件的语义得到了其它汉语表达式的恰当注解,该汉字本身的语义也便可得到确认。在本文中,许慎的"六书"论将以一种可计算的方式而得到系统化的重构,而该重构的技术基础则是王培先生发明的"纳思系统"(非公理化推理系统)。在这种重构基础上,现代汉语中语词的语义复合方式,也可以得到一种新颖的理解。
As a critic of Strong AI, John Searle claims that a digital computer processing Chi- nese inputs can never understand Chinese as human Chinese speakers do, since it can never acquire the semantic content of Chinese whereas the latter can. His argument as- sumes both a semantics-syntax dichotomy and an externalist point of view of semantics itself. I will try to meet Searle's challenge by denying both assumptions. What I will take is an internalist-and-quasi-holistic approach to semantics, according to which meanings are brought into a semantic network without bearing relationships with external entities which cannot be internalized in the cognitive system. Moreover, the intended seman- tic model is not syntactically driven, since there is no content-neutral rule for forming compound meanings out of more "atomic" ones. Rather, the affinity between "atomic" meanings itself plays a pivotal role in forming compound meanings. Moreover, the se- mantic issue of how the fusion of meanings is possible is not privileged by contemporary linguists or philosophers of language. It was addressed by the ancient Chinese lexicog- rapher Shen Xu (ca. 58 - ca. 147) in his "Six Writings" theory, according to which the meaning of a Chinese character can be recognized if the semantic status of its components can be glossed by other Chinese expressions the meanings of which are accessible to the reader. I will try to reconstruct Xu's theory in a computable manner with the aid of Non- Axiomatic Reasoning System, and the resulting model can also help us in understanding how the words are constituted by characters in modern Chinese.
出处
《逻辑学研究》
CSSCI
2012年第2期1-49,共49页
Studies in Logic
基金
霍英东教育基金会第12届青年教师资助项目(NO.121095)
复旦大学"985工程"三期整体推进人文学科项目(NO.2011RWXKYB038)
全国优秀博士论文后续资助项目(NO.200701)的资助