Predicate-Argument Structure Divergences in Chinese and English Parallel Sentences and their Impact on Language Transfer

2025-11-12

AI Summary

This study investigates how predicate-argument structure (PAS) differences between Chinese and English affect cross-lingual transfer. To address this, we propose the first systematic typology of Chinese–English PAS divergences, grounded in parallel sentence pairs. Leveraging annotation projection experiments and multilingual pretrained language models, we quantitatively analyze structural alignment and misalignment. Results reveal pronounced asymmetry in cross-lingual transfer: argument structure projection from English to Chinese significantly outperforms the reverse direction. This finding provides the first empirical evidence—grounded specifically in predicate-argument syntax—of an intrinsic asymmetry in cross-lingual transfer mechanisms. It advances theoretical understanding of transfer phenomena and yields actionable insights for source-language selection, cross-lingual dependency parsing, and NLP in low-resource languages.

Technology Category

• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Cognitive Modeling & Cognitive Systems: Analogy
• Knowledge Representation and Reasoning: Argumentation

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• User Modeling, Personalization and Recommendation: Practical large-scale studies of user experience

Abstract

Cross-lingual Natural Language Processing (NLP) has gained significant traction in recent years, offering practical solutions in low-resource settings by transferring linguistic knowledge from resource-rich to low-resource languages. This field leverages techniques like annotation projection and model transfer for language adaptation, supported by multilingual pre-trained language models. However, linguistic divergences hinder language transfer, especially among typologically distant languages. In this paper, we present an analysis of predicate-argument structures in parallel Chinese and English sentences. We explore the alignment and misalignment of predicate annotations, inspecting similarities and differences and proposing a categorization of structural divergences. The analysis and the categorization are supported by a qualitative and quantitative analysis of the results of an annotation projection experiment, in which, in turn, one of the two languages has been used as source language to project annotations into the corresponding parallel sentences. The results of this analysis show clearly that language transfer is asymmetric. An aspect that requires attention when it comes to selecting the source language in transfer learning applications and that needs to be investigated before any scientific claim about cross-lingual NLP is proposed.

Problem

• Analyzing predicate-argument structure divergences in Chinese-English parallel sentences
• Investigating asymmetric language transfer effects in cross-lingual NLP
• Categorizing structural divergences impacting annotation projection between languages

Innovation

• Analyzes predicate-argument structure divergences in parallel sentences
• Proposes categorization of structural divergences through annotation projection
• Reveals asymmetric language transfer in cross-lingual NLP applications