Construction and application of a multimodal interlanguage corpus for teaching chinese as a foreign language

This thesis focuses on the construction and application of a pedagogically oriented multimodal Chinese interlanguage corpus, MICICL (Multimodal Interlanguage Corpus for Italian Chinese Learners). The corpus targets the language production of Italian L1 learners of Chinese and systematically integrates handwritten and reading-aloud data. Through a unified data processing and annotation framework, learner outputs from different modalities are made directly comparable within a single analytical structure. MICICL comprises language data from 39 learners collected across multiple tasks, including 155 handwritten samples and 153 corresponding reading-aloud samples. For each data unit, the original media files, transcriptions, error annotations, and minimally edited corrected versions. This design results in a structured, searchable, and reusable dataset. Given the formal and procedural differences between handwritten and reading-aloud data, separate transcription, annotation, and correction protocols were developed for each modality, with an emphasis on operational feasibility and pedagogical interpretability. To illustrate the application of the corpus, the thesis presents a character-level analysis as an example of how MICICL can be used in interlanguage research and pedagogical analysis. The illustrative analyses include quantitative descriptions of character-level phenomena in handwriting, phonological performance in reading-aloud data, writing-reading correspondences, and comparisons across learners at different proficiency levels. These analyses are intended to demonstrate the corpus’s capacity to support within-modality and cross-modality analysis and to provide data-based support for teaching-related research and practice. Overall, through the construction of MICICL and its application examples, this thesis presents a workflow for building and using a multimodal Chinese interlanguage corpus, from data collection and processing to analytical implementation. It offers a practical reference for the development of similar corpus resources and their use in both research and pedagogical contexts.