China’s Security Signaling in the Taiwan Strait: Narratives and Patterns (2016–2024)

Abstract This paper examines patterns in geo-strategic narratives associated with China's security signaling in the Taiwan Strait since 2020, showing that variations in China’s pressure-oriented diplomatic messaging are associated not only with reactions to U.S. or Taiwanese actions, but also with Beijing’s reassessment of global power dynamics—framed as the “great changes unseen in a century”—and growing confidence in its national capabilities. To test this argument, the paper employs a large language model (LLM) alongside advanced computational techniques to perform a longitudinal content analysis of nearly 2,000 People’s Daily articles from 2016 to 2024. The study classifies articles using Google’s Gemini 2.0 Flash model, and employs statistical modeling to examine how narrative indicators that capture geo-strategic reassessment and nationalist aspirations are associated with variations in Beijing’s security signaling toward Taiwan. Results indicate that China’s assertive messaging toward Taiwan is strongly associated with nationalist, identity-based narratives, while its coercive signaling shows a stronger correlation with China’s geo-strategic recalculation. Importantly, the findings also reveal that China's geo-strategic narratives are not monologic: different components of the “great changes unseen in a century” narrative are associated with distinct—and in some cases contradictory—patterns in Beijing’s coercive posture in the Taiwan Strait. This finding challenges the prevailing accounts of Chinese revisionism. The study contributes to a more sophisticated reading of Chinese security signaling, with broader implications for the management of regional stability in the West Pacific. Theoretically, the study advances the understanding of Chinese revisionism by disaggregating Beijing's “great changes unseen in a century” framework into distinct narrative components with divergent policy implications. Methodologically, it demonstrates the application of large language model classification combined with retrieval-augmented generation (RAG) as a validated instrument for longitudinal content analysis in social science research.