Liaozhai through the Looking-Glass: On Paratextual Explicitation of Culture-Bound Terms in Machine Translation

2025-09-27

AI Summary

Machine translation (MT) remains limited in conveying the deep contextual meanings of culture-loaded terms, as current approaches focus exclusively on intra-textual processing and overlook paratextual strategies—such as footnotes and endnotes—commonly employed by human translators. To address this gap, this work formally integrates Genette’s paratextual theory into MT for the first time, introducing the novel task of *paratextual explicitation*. We construct a high-quality dataset of 560 expert-aligned paratextual annotations. Leveraging large language models (LLMs), we generate paratexts via reasoning trajectory analysis, intrinsic prompting, and agent-based retrieval. Human evaluation confirms that LLM-generated paratexts significantly enhance readers’ cultural comprehension—though still falling short of professional human performance—thereby validating technical feasibility. Moreover, our analysis uncovers systematic differences in expert translators’ paratextual strategies, paving the way for scalable, culturally adaptive, personalized explanations and cross-cultural communication.

Technology Category

• Natural Language Processing: Machine Translation, Multilinguality, Cross-Lingual NLP
• Machine Learning: Transparent, Interpretable, Explainable ML
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

The faithful transfer of contextually-embedded meaning continues to challenge contemporary machine translation (MT), particularly in the rendering of culture-bound terms--expressions or concepts rooted in specific languages or cultures, resisting direct linguistic transfer. Existing computational approaches to explicitating these terms have focused exclusively on in-text solutions, overlooking paratextual apparatus in the footnotes and endnotes employed by professional translators. In this paper, we formalize Genette's (1987) theory of paratexts from literary and translation studies to introduce the task of paratextual explicitation for MT. We construct a dataset of 560 expert-aligned paratexts from four English translations of the classical Chinese short story collection Liaozhai and evaluate LLMs with and without reasoning traces on choice and content of explicitation. Experiments across intrinsic prompting and agentic retrieval methods establish the difficulty of this task, with human evaluation showing that LLM-generated paratexts improve audience comprehension, though remain considerably less effective than translator-authored ones. Beyond model performance, statistical analysis reveals that even professional translators vary widely in their use of paratexts, suggesting that cultural mediation is inherently open-ended rather than prescriptive. Our findings demonstrate the potential of paratextual explicitation in advancing MT beyond linguistic equivalence, with promising extensions to monolingual explanation and personalized adaptation.

Problem

• Addressing cultural term translation challenges in machine translation systems
• Introducing paratextual explicitation using footnotes for cultural mediation
• Evaluating LLM performance on cultural explanation generation tasks

Innovation

• Formalizing paratext theory for machine translation tasks
• Evaluating LLMs with reasoning traces on paratext generation
• Using statistical analysis to reveal translator variation patterns