Pragmatics and causal connectives: A contrastive study of Korean and English using the AI-Hub parallel corpus

This interdisciplinary study analyzes Korean and English causal connective expressions using the AI-Hub Korean-English parallel corpus. The primary objective is to identify the unique linguistic and cultural features of Korean causal connectives by comparing them with their English counterparts. Korean includes a wide range of causal connectives, many of which exhibit additional pragmatic features such as +negative, +uncertainty, and +plurality. From both linguistic and cultural perspectives, this study investigates whether these features are exclusive to Korean and explores the cultural factors contributing to their prevalence. To extend the analysis into a computational framework, the study defines a formal task for evaluating the preservation of pragmatic meaning in translation. Specifically, each Korean-English sentence pair is annotated for the pragmatic features expressed in Korean, and the extent to which those features are retained in the English translation is assessed. Two task formulations are proposed: (1) a binary classification indicating full preservation vs. loss or shift, and (2) a continuous ?pragmatic shift score? ranging from 0.0 to 1.0. This enables future implementation of rule based or learning-based models to detect pragmatic mismatches in translation. The analysis of seven Korean causal connectives reveals that the additional pragmatic features are specific to Korean and rarely appear in English. These features are culturally grounded: the +negative feature aligns with Korean speakers? tendency to express disapproval indirectly to preserve politeness; +uncertainty reflects a cultural preference for hedging and softening assertions; and +plurality indicates an avoidance of definitive statements, consistent with indirect and euphemistic communication strategies common in Korean discourse. Ethical considerations regarding data licensing and cultural bias are addressed. This research offers practical implications for computational linguistics, translation studies, and Korean language education. By uncovering culturally embedded differences in how causality is expressed, the study enhances cross-cultural understanding and contributes to improved communication in multilingual contexts.