Statistical Learning of L2 Chinese Characters: The Role of Phonological Similarity and Explicit Instruction

Learning to read in Chinese requires learners to map phonological forms not only to meaning but also to visually complex logographic characters. The present study examined whether native English speakers can acquire novel Chinese characters through cross-situational statistical learning (CSSL), and whether phonological similarity constrains this process. Fifty-eight English-speaking learners completed a CSSL task in which they learned associations between spoken Chinese pseudowords and Chinese characters without explicit feedback. Phonological similarity was manipulated such that some character pairs were associated with tonal minimal pairs, whereas others involved phonologically distinct words. In addition, half of the participants received brief explicit instruction on Chinese lexical tones prior to the task. Results showed that learners successfully acquired character–sound mappings after a short exposure period, demonstrating that statistical learning mechanisms can support early stages of Chinese character acquisition. However, learning was significantly constrained by phonological similarity, with tonal minimal pairs proving especially difficult. No significant differences emerged between the instructed and implicit learning groups. Nevertheless, post-task questionnaires revealed that participants who reported greater awareness of lexical tones achieved higher overall accuracy. These findings highlight the importance of phonological distinctiveness in character learning and suggest that learners’ sensitivity to tonal information may facilitate acquisition even in predominantly implicit learning contexts. The study contributes to current research on the integration of phonological and orthographic learning in Chinese second language acquisition.