This section introduces the creation background, philosophy, process, and value of the jaGLOBE Contemporary Japanese Written Language Balanced Corpus. As the Japanese sub-corpus of the "BFSU Global Corpus of Languages (GLOBE)", jaGLOBE collects various genres of Japanese written texts from 2013 to 2023. It follows the Brown Corpus sampling principles to build a million-word-level balanced corpus resource, supports multiple search methods, and provides a fundamental resource for Japanese language research, teaching, and multilingual comparative studies.

1 The History and Current State of Japanese Corpus Construction in Japan and China

This section details the history and current state of Japanese corpus construction in Japan and China. Since the 1990s, with the widespread adoption of computer technology and the development of natural language processing, Japan began constructing electronic general native language corpora. The Kyoto University Text Corpus was launched in 1997, and the field entered a systematic development phase after 2000. From 2006 to 2010, the National Institute for Japanese Language and Linguistics developed the large-scale, influential Balanced Corpus of Contemporary Written Japanese (BCCWJ) and created the NLB intelligent search platform. Additionally, specialized corpora such as the Corpus of Spontaneous Japanese (CSJ) and learner corpora were established. In China, the Chinese-Japanese Parallel Corpus was released in 2003, followed by a nearly decade-long period of relative inactivity. The Chinese Japanese Learner Corpus and the Japanese Learner Written Corpus were released in 2012 and 2020, respectively. Corpus construction in both countries has provided data resources for empirical research in related fields, but issues such as data obsolescence and imbalance persist. The creation of the jaGLOBE corpus aims to address these shortcomings and provide technical support for comparative studies on typological features across multiple languages.

2 System Construction and Technical Features of the jaGLOBE Corpus

This section elaborates on the system construction and technical features of the jaGLOBE corpus. jaGLOBE is a million-word-level balanced corpus created based on the Brown Corpus model, comprising 500 texts, each approximately 2000 words long. The corpus is divided into two major genres: Informative and Imaginative, covering various sub-genre types, and has been adjusted according to Japanese cultural characteristics. The construction process includes text sampling, processing, entry, and metadata annotation. It strictly selects original texts authored by native Japanese speakers, excluding translated texts. After preprocessing, word segmentation, and part-of-speech tagging, the final corpus consists of 1,149,414 words. jaGLOBE features three main characteristics: timeliness, balance, and practicality. All texts are sourced from after 2010. Stratified sampling ensures broad and balanced source representation. Hosted on the CQPweb platform, it supports various complex query needs, providing effective support for language contrastive research and foreign language teaching.

3 Current Status and Future Prospects of Applied Research Using Japanese Corpora

3

First, it reviews the current state of applied research using Japanese corpora in Japan and China from 2003 to 2023. Bibliometric analysis reveals that Japanese corpora have become a hotspot in supporting acquisition research within the field of applied linguistics and are prominently used in theoretical linguistic research areas such as lexicology and semantics. It then introduces two research paradigms: corpus-based research and corpus-driven research, noting that current Japanese corpus research predominantly follows the former. Subsequently, it lists various specialized corpora and their outcomes recently constructed by researchers in Japan and China, such as the "Daichi Corpus" and the "Japanese Interlanguage Speech Corpus". Finally, looking to the future, it suggests that Japanese corpora will develop towards multimodality, dynamism, and intelligence, thereby broadening the boundaries of Japanese linguistic research.

4 Conclusion

It emphasizes the significant value of the jaGLOBE corpus as the Japanese sub-corpus of the "BFSU Global Corpus of Languages". Not only does it provide data support for Japanese language ontology research covering the past decade, but by following the Brown Corpus model and forming an alliance with multilingual sub-corpora, it has also constructed a digital platform supporting multilingual comparative studies. This enables researchers to conduct independent research or cross-linguistic comparative analysis, promoting an upgrade in linguistic research paradigms.