Abstract
<p>This study investigates how pitch gesture (PG)—co-speech hand movements that iconically trace the contour of Mandarin tone—works in a real online beginner Chinese as a foreign language (CFL) class. Three sessions were recorded and annotated using ELAN 7.0 (Max Planck Institute for Psycholinguistics, 2025). Learner participation varied across sessions: one learner attended only the first session, while two learners completed all three sessions. The study addressed three questions: what happens to learner tonal accuracy when the teacher shifts from PG to semantic gestures within the same lesson using the same vocabulary; whether learners internalize PG as a self-correction tool when teacher PG support is reduced; and whether learner PG use is related to higher tonal accuracy across sessions.The results demonstrate a consistent pattern across all three sessions: In session 3, which used a four-phase design with the same vocabulary throughout, tonal accuracy fell from 95% (S2) and 76% (S3) in phase 1 (full PG support) to 43% and 25% in Phase 4 (semantic gestures only), a drop of around 51-52 percentage points for both learners, despite their different L1 backgrounds.These findings suggest that PG and semantic gestures serve different cognitive functions in tonal language learning: PG appears to support tonal production by aligning visual and phonological input, while iconic gestures and metaphoric gestures are relevant to a sharp drop in tonal accuracy even when used with familiar vocabulary. This study provides naturalistic classroom evidence that the interference effect of semantic gestures on tonal processing also manifests in spoken production and points to the importance of sequencing gesture type carefully in tonal language pedagogy.</p>