摘要
人工智能在自然语言处理领域目前已取得显著进展,然而其在文学批评中的有效性与适用性仍存在学理争议。从技术层面来看,人工智能主要依赖于分词与词向量等语言表征方式,尽管这些方法在表层语义分析上表现出一定的效能,但在深层次的语义理解方面仍面临系统性的认知障碍。不仅如此,人工智能的生成机制虽能有效模拟自然语言中的物理规律,但由于缺乏具身认知所依赖的感知经验,难以实现从具体经验到审美判断的认知跃迁。此外,虽然强化学习和思维链技术显著提升了其执行既定任务的能力,但人工智能无法实现自我审视和价值重构,亦无法进行伦理判断与历史反思,限制了创造性思维的生成。文学批评依赖批评主体的价值判断、历史意识与多重对话能力,因此,人工智能的工具理性本质与文学批评的价值理性诉求存在根本冲突,这就决定了人工智能尚且无法替代人类批评家的主体性洞察与意义诠释。
Artificial intelligence has made significant progress in the field of natural language processing,yet its applicability in literary criticism remains controversial.The language representation approaches reliant on tokenization and embedding techniques exhibit inherent limitations in semantic understanding.Although its generation mechanism,grounded in mathematical probabilities,can simulate physical laws,AI struggles to achieve the cognitive leap from concrete experience to aesthetic judgment due to the lack of embodied experience.Furthermore,the existing technical architecture fundamentally limits the generation of reflective and creative thinking.Literary criticism involves not only text analysis but also relies on the critic's capacity for value judgment,emotional empathy,and historical interpretation.Therefore,the rational nature of AI tools fundamentally conflicts with the value-driven demands of literary criticism,and this essential difference determines that AI cannot replace the subjective insight and meaning interpretation of human critics.
作者
陈琴
陈亚丽
Chen Qin;Chen Yali
出处
《浙江学刊》
北大核心
2025年第3期97-105,239,共10页
Zhejiang Academic Journal
作者简介
陈琴,首都师范大学文学院博士研究生;陈亚丽,首都师范大学文学院教授、博士导师。
