对“汉字屋论证”逻辑结构的五种诊断模式  

Five Diagnoses of the Logical Structure of Searle's Chinese Room Argument

作　　者：徐英瑾[1] 

机构地区：[1]复旦大学哲学学院,上海200433

出　　处：《复旦学报（社会科学版）》2008年第3期82-89,共8页Fudan Journal（Social Sciences）

基　　金：2006年国家社科基金项目"维特根斯坦哲学视野中的人工智能问题"(项目批准号:06CZX011)的资助

摘　　要：美国哲学家J.R.塞尔(Searle)于1980年提出的"汉字屋论证"(Chinese Room Argument,以下简称为CRA)目前早已成为人工智能哲学的经典话题之一。二十七年以来,反驳CRA的学者往往致力于揭露CRA本身的形而上学前设的错误,却不太去检讨CRA本身的逻辑形式。而美国学者郝泽(Hauser)与英国学者丹普尔(Damper)虽然就CRA的逻辑形式提出了总计四种诊断模式,却都没有正确地揭示出CRA的真正错误。本文将在他们的基础上提出对于CRA逻辑结构的第五种诊断模式,并由此论证:CRA之所以失败,乃是因为塞尔并未根据该论证的内在要求而在"汉字屋系统"与"计算机系统"之间建立起一种恰当的同构关系来。Since Searle＇s paper “Mind, Brain and Programs” was published in 1980, his Chinese Room Argument （CRA）, which is intended for arguing against the possibility of Strong Artificial Intelligence, has long been the target of counter-attackers who believe that CRA is wrong. Unfortunately, standard replies to CRA typically focus on the metaphysical presuppositions involved in the argument rather than dig out its logical structure. Although Hauser and Damper used to evaluate CRA from a logical point of view by offering four relevant diagnoses, their bullets missed the target, either. The fifth diagnosis of the logic of CRA, presented by this paper, will demonstrate the invalidity of CRA by debunking Searle＇s failure to establish a proper isomorphism between a Chinese Room System and a Computer System-an isomorphism without which CRA cannot make any sense,

关 键 词：强人工智能 汉字屋论证 句法学 语义学 思想实验 同构关系 

分 类 号：B812[哲学宗教—逻辑学]