The Classic of Mountains is part of the pre-Qin ancient text The Classic of Mountains and Seas (Shanhaijing 山海经). Within approximately 2,000 characters, it records a vast array of mountains, rivers, flora and fauna, and minerals in a highly stylized narrative, making it the earliest extant geographical and ecological compendium in China. However, research on The Classic of Mountains has been confined to traditional humanities research paradigms, and the related knowledge has not yet been systematically digitized.
This study attempts to leverage artificial intelligence to collect, integrate, and present historical knowledge in The Classic of Mountains, thereby exploring how and to what extent AI can contribute to humanities research. It aims to investigate the geographical and ecological landscapes recorded in the text by constructing a commentary database and a knowledge-integration platform using technical methods such as prompt engineering, pre-trained language models, GraphRAG, geocoding, and boundary estimation algorithms.
The visualizations presented in this project offer a new way to understand ancient peoples’ perception and conceptualization of space and environment and may also shed light on the long-debated question of the text’s geographical and ecological authenticity. The knowledge-integration platform and commentary database developed in this project can support humanities scholars in their research and serve as a digital practice of traditional philological exegesis (xun gu 訓詁) concept of collected commentary (ji shi 集釋).
Knowledge graph of the three most biodiverse mountains in The Classic of Mountains. Image by Ke Liang.