Introduction
Zoumalou Slips
Figures of Zoumalou Slips ( shown above )
https://www.chinajiandu.cn/Collection/List/wj
Zoumalou Slips are a large number of bamboo slips dating back to the Wu Kingdom during the Three Kingdoms period. In 1996, there were found in an ancient well numbered J22 under the foundation of the former No. 50 Zouma Street in the center of Changsha City, and that is how it derived the name.
The Database Involved
2,141 bamboo slip records
The textual information from this portion of slips has been compiled into a database by Professor CHEUNG Kwong Yue Alex and Professor LAI Ming Chiu from CUHK, and further edited by Master LAU Ting Long. It is stored on the website of our university’s History Department and serves as the data source for our research. The link is as follows. https://www.history.cuhk.edu.hk/tc/elkt/tianjiabie/
Historical Background Narrative
The Three Kingdoms period was a turbulent era in Chinese history. Due to continuous warfare, the population suffered severe decline, and the agricultural economy was heavily disrupted. After the establishment of the Wu kingdom, the government implemented a land rental system on scattered state-owned lands to restore production. Changsha County serves as an example of such implementation.
In the bamboo slips, we can observe the following information: recorded dates (down to the day), specific locations (down to the villages under Changsha County), names and identities of tenants, types and areas of rented fields, drought conditions, rental payments, etc.
Synopsis of our work
The data mining of the text translation of the LLM model makes the visualizations organically connected, so that readers can better understand from the charts the causes of the changes in the field tax and money and the implied historical origins against the translated text. This provides a deeper impression of the times as far back as thousands of years ago.
Team Members
LIU Leyi, Lyra (CSCIN, Year 2)
MA Heyang, Martin (CSCIN, Year 2)
Michael YIP (Coach)
Vincent LUM (Coach)
Acknowledgement
長沙簡牘博物館,典藏,吳簡 https://www.chinajiandu.cn/Collection/List/wj
蒋福亚. (2012). 走马楼吴简经济文书研究 (第1版.). 国家图书馆出版社.
沈刚. (2017). 《长沙走马楼三国吴简》语词汇释 = A compiling of the word explanations in Wu State’s wooden and bamboo slip documents unearthed in Zoumalou Changsha City (第1版.). 中国社会科学出版社.
張光裕, & 黎明釗. (n.d.). 《長沙走馬樓三國吳簡‧嘉禾吏民田家莂》全文檢索資料庫. (劉天朗, Ed.). Chinese University of Hong Kong Department of History. https://www.history.cuhk.edu.hk/tc/elkt/tianjiabie/
Deep learning-based methods for natural hazard named entity recognition – Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/The-overall-structure-of-the-BERT-model_fig1_359301499 [accessed 11 Mar, 2024]
Guo, G., Yang, J., Lu, F., Qin, J., Tang, T., & Zhao, W. X. (2023). Towards Effective Ancient Chinese Translation: Dataset, Model, and Evaluation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 14303, pp. 416–427). https://doi.org/10.1007/978-3-031-44696-2_33
Wikipedia – Zoumalou bamboo slips http://en.wikipedia.org/wiki/Zoumalou_bamboo_slips
Data processing tools :
- Excel, Python : pandas, chinese2digits
Translation tools :
- OpenAI. (2023). ChatGPT (3.5-turbo) [Large language model]. https://chat.openai.com/chat
- Python package: CPT, a Pre-Trained Unbalanced Transformer for Both Chinese Language Understanding and Generation. https://github.com/fastnlp/CPT/tree/master
- Python package: Erya-4FT, a pretrained model specifically designed for translating Ancient Chinese into Modern Chinese. https://github.com/RUCAIBox/Erya
- Python package: “Jieba” (Chinese for “to stutter”) Chinese text segmentation. https://github.com/fxsjy/jieba
- Python package: OpenCC, an opensource project for conversions between Traditional Chinese, Simplified Chinese and Japanese Kanji (Shinjitai). https://github.com/BYVoid/OpenCC
Visualization tools :
- Excel, Python : matplotlib, seaborn