TEAL Digital Scholarship Series 2025-26: Detecting Shifts in Linguistic Register in Late Imperial Chinese Fiction

The Tateuchi East Asia Library (TEAL) is proud to present the 2025-2026 TEAL Digital Scholarship Series, a dynamic program showcasing cutting-edge research by scholars in the fields of Chinese, Japanese and Korean studies. This series highlights how innovative digital tools and methodologies are transforming East Asian scholarship, fostering interdisciplinary dialogue, and broadening the impact of research within and beyond academia.
 

Detecting shifts in linguistic register in late imperial Chinese fiction: Fine-tuning language models to detect fictionalized memorials to the emperor

Abstract:

It is common in late imperial Chinese literature for novels to appropriate the voice of officialdom for a variety of purposes, often as a means of bolstering historical credibility. While this appropriation can manifest in a variety of ways, it often comes in the form of verbatim quotations from memorials that officials wrote to the emperor. Some such quotations are attested to in historical sources, but often the source of the memorial is either no longer extant or never existed in the first place. Understanding how memorials are deployed across fiction at scale can offer new insights into the role the appropriation of officialese plays in constructing historical narrative in late imperial fiction. While older intertextuality algorithms can help us identify memorials found in digitized historical sources, they cannot identify lost or inauthentic documents. In this talk based on part of his first monograph project, Paul Vierthaler will discuss experiments he has been conducting in leveraging the fundamentally distinct linguistic registers of fiction and memorials to fine-tune language models that can flexibly identify memorial-like language across hundreds of novels.

Bio:

Paul Vierthaler is an assistant professor of late imperial Chinese literature and interdisciplinary data science in the East Asian Studies department at Princeton University. He previously taught Chinese studies at the College of William & Mary and digital humanities at Leiden University, where he helped establish the Leiden University Centre for Digital Humanities. His research lies at the intersection of Ming and Qing literature, print history, and computational analytics. His first book project is a computational study of unofficial historical narratives found within fictional and unofficial historical genres of text written in late Ming and Qing China.