基于问卷调查的社会结构定量测量存在时间跨度有限、测量维度单一、隐性指标不足等问题。为此本文阐述了一种基于文本大数据，运用自然语言处理算法来测量不同时期的话语结构，进而反映社会结构及其变迁规律的方法。以中国社会职业地位的历史变迁（1940—2015）为例，本文基于书籍大数据，从财富、权力、文化、声望四个维度刻画了职业地位和职业地位结构的历史变迁图景。该方法对传统问卷测量方法形成了重要的补充，为缺乏问卷资料的场景，特别是大历史跨度下的主观观念结构和客观社会结构变迁的测量提供了新的计算社会学工具。

Traditional survey-based approaches to measuring social structure face several limitations, including restricted historical coverage, narrow measurement dimensions, and insufficient latent indicators. To address these challenges, this paper introduces a method that utilizes textual big data and natural language processing algorithms to analyze discourse structures across different periods, thereby reflecting social structure and its patterns of change. Taking the historical evolution of social and occupational status in China as an example, the research utilizes book-based textual big data to trace the historical transformation of occupational status and the structure of occupational status in China across four dimensions: wealth, power, education, and prestige. This method serves as an important complement to traditional survey-based measurement approaches, providing a new computational sociology tool for situations where survey data is unavailable, particularly for measuring changes in subjective perception structures and objective social structures over long historical periods.

本文系首批教育部哲学社会科学创新团队（南京大学“中国式现代化的社会治理数智研究创新团队”）成果。

欢迎阅读《社会学研究》！您是该文第293位读者！