TEI/XMLとRDFによる日本古辞書の出典情報記述モデル = A Description Model for Citation Information in Early Japanese Dictionaries using TEI/XML and RDF

다국어 초록 (Multilingual Abstract)

This paper proposes a new model for digitally structuring the complex citation information within the Zushoryō-bon Ruijū Myōgishō (ZRM), an 12th-century Japanese dictionary. The ZRM cites over 300 documents using unique abbreviations, posing a challenge for machine readability that conventional RDB approaches cannot adequately address.
This study presents a hybrid model linking two technologies. First, the main text’s logical structure (entries, definitions, readings, citations) is marked up as a hierarchical TEI/XML tree structure. Second, the extensive information on the 300+ source documents is separated from the text and constructed as a knowledge graph using the Resource Description Framework (RDF).
This graph-based RDF approach is essential for flexibly modeling complex philological relationships (e.g., one-to-many, many-to-many, and indirect “second-hand” citations). Furthermore, the citation ontology is designed as Linked Open Data (LOD), linking to external authorities like VIAF and Wikidata to ensure high reusability and interoperability with other major digital archives (e.g., SAT database).
In conclusion, this model links the TEI body to the external RDF ontology, achieving both data accuracy/reproducibility (TEI) and data flexibility/extensibility (RDF). This research provides a new foundation for philological study, moving beyond a single dictionary to enable the construction of a “Japanese Classical Knowledge Graph” visualizing the knowledge network of 12th-century Japan.