Eyes That Tell: Universal and Language-Specific Eye Movement Patterns in Chinese and Korean Speech P

This study used an eye-tracking visual-world paradigm to examine how speakers of different languages allocate visual attention during spoken event description. Thirty-six native Chinese and thirty-six native Korean speakers viewed transitive scenes (e.g., “The frog steps on the rabbit”) and described them aloud while their eye movements were recorded for three predefined regions of interest (agent, patient, and action). Two primary patterns emerged. First, regardless of language background, participants directed their earliest fixations to the action region shortly after scene onset, indicating rapid conceptual encoding of event relations before linguistic formulation. Second, only Korean speakers showed sustained regressions to the action region after speech onset. This post-vocalization return disappeared when the target verb was provided in advance, revealing a post-articulatory monitoring process linked to the verb-final, morphologically rich structure of Korean. These results demonstrate that language-specific morphosyntactic demands can modulate real-time visual attention during speech production, providing process-level evidence for a weak form of linguistic relativity.