Cross-linguistic adaptation of feature-cue-based speech labeling: The case of Korean (L).

Abstract

This study examines the cross-linguistic adaptation of a feature-cue-based model for speech labeling, initially developed for English, to the phonetic and phonological characteristics of Korean. Using a Korean speech corpus, this study applies the linguistic event extraction and interpretation framework to annotate individual acoustic cues associated with phonological features. The findings suggest that while the model's core principles broadly capture many phonetic details in Korean, certain adjustments are required to account for language-specific phonological distinctions and their phonetic signaling. This research enhances understanding of language-specific phonetic patterns within a universal framework and demonstrates the model's adaptability across languages.