休谟难题的融贯论解决方案  

A Solution of Coherence Theory to Hume's Problem

作　　者：梁贤华 LIANG Xianhua(School of Philosophy and Social Development,South China Normal University,Guangzhou,Guangdong,510631)

机构地区：[1]华南师范大学哲学与社会发展学院,广东广州510631

出　　处：《自然辩证法通讯》2023年第6期37-44,共8页Journal of Dialectics of Nature

基　　金：国家社科基金重大项目“逻辑真理论的历史源流、理论前沿与应用研究”(项目编号:17ZDA025)。

摘　　要：归纳推理的基石“将来和过去相似”,可以从贝叶斯决策论找到一个理论的基础:尽管不同的主体对同一个事件的评估(即主观概率赋值)很可能是不同的,但是,为了避免陷入必输无疑的荷兰赌困境,就同一个主体而言,他/她在进行概率赋值时应当保持融贯,即有必要假定将来和过去是相似的。这个方案强调的是形式的分析,因此可能被批评缺乏对经验事实的关注。所以,我们提出了一个形式和内容并重的休谟难题的解决方案,并且进一步提出了一种新的归纳推理研究方法:基于证据的贝叶斯主义进路。The cornerstone of inductive reasoning that “the future is similar to the past” can find a subjective basis in Bayes'decision theory,that is,although the evaluations(i.e.subjective probability assignment)of different subjects on the same thing are likely to be different,the same subject,as far as he/she is concerned,should remain accommodating in the conduct of probability assignment,i.e.assume that the future is similar to the past,so as to avoid falling into the Dutch Book dilemma which means necessary lost.However,this scheme emphasizes formal analysis,so it may be criticized for lack of attention to empirical facts.Therefore,we propose a solution to Hume's problem with equal emphasis on form and content,and further propose a new research method of inductive reasoning:an evidence-based Bayesian approach.

关 键 词：休谟难题 融贯 证据 

分 类 号：N031[自然科学总论—科学技术哲学] O212.8[理学—概率论与数理统计]