In response to changes in the educational environment, namely the advent of the artificial intelligence (AI) era and the downsizing of language and literature majors, this paper discusses the construction and utilization of a small-scale textbook corpus integrating Data-driven Learning (DDL) into Chinese language education. This study posits that to enhance the density of major courses—which has diminished due to the expansion of multi-major programs—and to strengthen technological convergence capabilities, a small-scale textbook corpus tailored to learners' proficiency levels can be more pedagogically oriented than a large-scale, general-purpose corpus. As a practical implementation method, this study collected 39 volumes from 14 series of Chinese language textbooks published in South Korea, China, the United States, and elsewhere. By applying metadata tagging and parsing specialized for dialogues and reading passages, the unstructured text format (.txt) was refined into a semi-structured JSON format (.json). Python-based natural language processing (NLP) packages, including Jieba and Stanza, were utilized for word segmentation and part-of-speech (POS) tagging. Based on the finalized data, a DuckDB-based corpus database (.duckdb) was constructed to facilitate regular expression and SQL queries. To demonstrate the educational applications of the constructed database, this paper presents several examples: the inductive discovery of grammatical rules for separable verbs using KWIC (Key Word in Context) search, a comparative analysis of high-frequency vocabulary across textbooks from different countries using Word Lists and Word Clouds, and a contrastive analysis of POS tagging results between heterogeneous tokenizers. Ultimately, this study suggests that a step-by-step, selfdirected language exploration process mediated by a corpus can serve as a practical model for convergence education, organically combining traditional language learning with modern computational technology.

본 연구는 AI 시대와 어문전공의 관심 축소라는 교육 환경 변화에 대응하여, 중국 어 교육에 데이터 주도학습(DDL)의 개념을 접목한 소규모 교재코퍼스구축 및 활용 방안을 제안한다. 다전공 확대로 낮아진 수업 밀도를 높이고 융합 역량을 강화하기 위해, 학습자 맞춤형 소규모 코퍼스가 대규모 범용 코퍼스보다 교육적으로 유용할 수 있다. 코퍼스는 한･중･미 등지의 교재 14종 39권을 수집해 JSON 포맷으로 정제 하였고 Jieba와 Stanza 등 파이썬 NLP 패키지로 단어 분절 및 품사 태깅을 수행한 뒤 DuckDB를 기반으로 데이터베이스화하였다. 그에 기초하여, KWIC 검색을 통한 이합사 특징 발견, Word List/Cloud를 활용한 국가별 고빈도 어휘 비교, 이종 토크 나이저 대조 분석 등의 활용방안을 제시한다. 본 연구는 코퍼스 기반 언어 탐구가 전통 어문학과 전산 기술을 결합한 실천적 융합교육 모델이 될 수 있음을 제안한다.