한문 교육용 기초한자 1,800자의 2음절 한자어 결합망: 위치 선호도와 네트워크 위상 구조 = Disyllabic Sino-Korean Word Network of the 1,800 Basic Hanja for Classical Chinese Education: Positional Bias and Topological Structure

다국어 초록 (Multilingual Abstract)

This study examines, from a network-analytic perspective, how the 1,800 Basic Hanja prescribed for Hanmun education combine in disyllabic Sino-Korean headwords recorded in Korean dictionaries, and aims to describe quantitatively the structural characteristics of the combinatory relations these characters form. For this purpose, disyllabic Sino-Korean words with Hanja spellings were extracted from Urimalsaem and the Standard Korean Language Dictionary, and only those headwords whose two constituent characters both belonged to the Basic Hanja list were selected. Based on these data, a Hanja combination network consisting of 1,795 nodes and 76,359 directed edges was constructed, and its degree distribution, community structure, positional preference, centrality, and core-layer indicators were analyzed.
The results show that the Basic Hanja combination network is sparse, with a density of 0.024, while also displaying a heavy-tailed distribution in which combinations are concentrated around a small number of hub characters, a pattern consistent with the characteristics of a scale-free network. In addition, the combination of average path length and clustering coefficient confirms the presence of small-world properties. High-degree characters such as 大, 中, 山, 水, and 人 functioned as central hubs in the network, and k-core analysis identified a maximum core number of 59, with the 530 characters belonging to the 59-core forming the structural backbone of the network. The five modules detected by the Louvain algorithm were interpretively labeled as “administration, scholarship, and public affairs,” “space, nature, and directionality,” “state, emotion, and abstraction,” “persons, ethics, and virtues,” and “time and quantity,” based on the prominent combinatory tendencies within each cluster; these modules also showed distinct structural differences and non-negligible cross-module connections. Analysis of the positional preference index (P-index) further revealed an overall tendency toward balanced left-right distribution, while positional biases observed in certain characters and combinatory groups provided quantitative clues for discussing functional tendencies in the word-formation process.
These findings demonstrate that the Basic Hanja participate extensively in the combinatory system of disyllabic Sino-Korean words and form key points of word formation within a multilayered structure characterized by hubs, cores, modules, and positional preference. The study thus presents, in quantitative terms, the structural characteristics of the Basic Hanja combination network and provides quantitative and structural grounds that may inform the design of Hanja and Sino-Korean vocabulary education, including core-layer-based sequencing, the pedagogical use of balanced-type characters, and the presentation of vocabulary based on combinatory modules. It is also expected to serve as foundational material for future research on network-based Hanja/Hanmun education and digitally oriented Hanja studies.