ZiNet: Linking Chinese Characters Spanning Three Thousand Years

Findings of the Association for Computational Linguistics: ACL 2022, pages 3061 - 3070 May 22-27, 2022 c⃝2022 Association for Computational Linguistics ZiNet: Linking Chinese Characters Spanning Three Thousand Years Yang Chi1, Fausto Giunchiglia1,2,3, Daqian Shi3, Xiaolei Diao3 Chuntao Li4, Hao Xu1,2,∗ 1School of Artificial Intelligence, Jilin University, Changchun, China 2College of Computer Science and Technology, Jilin University, Changchun, China 3DISI, University of Trento, Trento, Italy 4School of Archaeology, Jilin University, Changchun, China yangchi19@mails.jlu.edu.cn, {xuhao,lct33}@jlu.edu.cn {fausto.giunchiglia,daqian.shi,xiaolei.diao}@unitn.it Abstract Modern Chinese characters evolved from 3,000 years ago. Up to now, tens of thousands of glyphs of ancient characters have been dis- covered, which must be deciphered by experts to interpret unearthed documents. Experts usu- ally need to compare each ancient character to be examined with similar known ones in whole historical periods. However, it is in- evitably limited by human memory and expe- rience, which often cost a lot of time but as- sociations are limited to a small scope. To help researchers discover glyph similar char- acters, this paper introduces ZiNet, the first di- achronic knowledge base describing relation- ships and evolution of Chinese characters and words. In addition, powered by the knowledge of radical systems in ZiNet, this paper intro- duces glyph similarity measurement between ancient Chinese characters, which could cap- ture similar glyph pairs that are potentially re- lated in origins or semantics. Results show strong positive correlations between scores from the method and from human experts. Fi- nally, qualitative analysis and implicit future applications are presented. 1 Introduction The evaluation of Chinese character can be divided into two stages: the ancient stage (before Han dy- nasty, 202 BC) and the clerical and standard script stage (after Han dynasty) (Qiu et al., 2000). At the former stage, ancient characters do not have a fixed shape, and their glyphs show several dif- ferences respect to modern characters: representa- tives include the Oracle bone script (Oracle) in the Shang Dynasty (about 1300 BC), which appears on animal bones or turtle shells (Boltz, 1986), the Chinese bronze script (Bronze, about 1000 BC) ap- peared on bronze wares (Shaughnessy, 1991) and the script belonging to the Warring States period (States), mainly recorded on wooden slips (about ∗Corresponding author Oracle bone script (~1300BC) Bronze script (~1000BC) Script in the Warring States period(~476BC) Small seal Script (~221BC) Clerical script (~202BC) Regular script (~220AD) Ancient characters 1 2 2 1 刀 knife 宿 stay overnight 牢 animal pen glyph1: glyph2: 宀+牛 宀+羊 radicals: 牛 (cow) 羊 (sheep) 宀 (house) glyph1: glyph2: 亻+㐁 宀+亻+㐁 radicals: 亻(person) 㐁 (mat) 宀 (house) Figure 1: Examples of historical evolution of Chinese characters associated with unfixed glyphs and radical compositions (the pictures on the top show unearthed ancient characters respectively written on the turtle shell, bronze ware and wooden slips). 400 BC) (Qiu, 2014). Evolution of glyphs of Chi- nese characters can be observed in Figure 1. Ancient unearthed documents show a wealth of information regarding that historical periods (Boltz, 1986; Shaughnessy, 1991; Qiu, 2014), which is great significant for understanding the …

Content truncated to 500 words