Word Embeddings Track Social Group Changes Across 70 Years in China

2025-04-12

AI Summary

This study investigates the diachronic evolution of semantic representations of social groups in Chinese official media texts (1950–2019), aiming to uncover how revolutionary societal transformations are linguistically encoded in state discourse. Leveraging a large-scale diachronic corpus, we apply multi-granular dynamic word embeddings (e.g., Dynamic Word2Vec, Temporal GloVe), semantic analogies and bias detection, and cross-temporal similarity trajectory analysis. To our knowledge, this is the first systematic characterization of social representation shifts within a non-Western, state-dominated discursive context. Results reveal high stability in ethnic, age, and body-size dimensions, whereas gender and class representations exhibit abrupt, event-driven shifts—strongly correlated with landmark historical developments (e.g., Reform and Opening-Up, women’s liberation)—challenging Western assumptions of gradual semantic evolution. The findings establish a critical non-Western benchmark for computational social science and empirically demonstrate dimension-specific heterogeneity and historical sensitivity in Chinese sociosemantic change.

Technology Category

• Natural Language Processing: Lexical Semantics and Morphology
• Application Domains: Humanities & Computational Social Science
• Knowledge Representation and Reasoning: Argumentation

Application Category

• Web Mining and Content Analysis: Models for Web evolution
• Social Networks and Social Media: Computational social science
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Language encodes societal beliefs about social groups through word patterns. While computational methods like word embeddings enable quantitative analysis of these patterns, studies have primarily examined gradual shifts in Western contexts. We present the first large-scale computational analysis of Chinese state-controlled media (1950-2019) to examine how revolutionary social transformations are reflected in official linguistic representations of social groups. Using diachronic word embeddings at multiple temporal resolutions, we find that Chinese representations differ significantly from Western counterparts, particularly regarding economic status, ethnicity, and gender. These representations show distinct evolutionary dynamics: while stereotypes of ethnicity, age, and body type remain remarkably stable across political upheavals, representations of gender and economic classes undergo dramatic shifts tracking historical transformations. This work advances our understanding of how officially sanctioned discourse encodes social structure through language while highlighting the importance of non-Western perspectives in computational social science.

Problem

• Analyze Chinese media's linguistic representation of social groups
• Compare Chinese and Western social group embedding patterns
• Track stability and shifts in stereotypes across political changes

Innovation

• Diachronic word embeddings analyze Chinese media
• Multi-temporal resolution tracks social group changes
• Non-Western perspectives reveal unique evolutionary dynamics