How to Build up a Computer which can Really Understand Chinese? —a “Six-Writings” Theory-Inspired Model of Processing Chinese Characters and Other Expressions

Yingjin Xu

DOI: https://doi.org/10.3969/j.issn.1674-3202.2012.02.001

2012-01-01

Abstract:As a critic of Strong AI, John Searle claims that a digital computer processing Chinese 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 assumes both a semantics-syntax dichotomy and an externalist point of view of semantics itself. IwilltrytomeetSearle’schallengebydenyingbothassumptions. WhatIwilltake 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 semantic 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 semantic 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 lexicographer Shen Xu (ca. 58 ca. 147) in his "Six Writings" theory, according to which the meaningofaChinesecharactercanberecognizedifthesemanticstatusofitscomponents 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 NonAxiomatic Reasoning System, and the resulting model can also help us in understanding how the words are constituted by characters in modern Chinese.