Korean Journal of Buddhist Studies ISSN:1598-0642(Print) 불교학연구

This paper is an interim report on the Tattvasaṅgrahapañjikā Encoding Project, which encodes Śāntarakṣita’s Tattvasaṅgraha and Kamalaśīla’s commentary, the Tattvasaṅgrahapañjikā, in TEI-XML. The project records the “reference information” generated when the two texts refer to external entities—persons, schools, views, and texts—in order to read the work as a whole from a distance and to reconstruct the eighth-century Indian intellectual network in which it is embedded. The paper first describes the work our team has carried out over the past two years: (1) preprocessing a legacy e-text keyed in the 1990s, by bringing it into conformity with Shastri’s edition and spacing it properly; (2) designing a TEI schema that records the polemical structure of the śāstra with <div>, referential terms with <rs>, and quotations, understood as explicit acts of attribution, with <q>, <seg>, and <rs> in combination; and (3) encoding the explicit reference information of the first nineteen chapters. It also presents a sample visualization of chapter 17, which offers a glimpse of what the encoded files will make possible. The paper then explores how agentic AI might transform the remaining work. Because interpretative judgment is inherent in every stage of the project, full automation is not feasible; yet collaboration with AI promises considerable gains in efficiency. Revisiting each stage—preprocessing, schema design, encoding, and visualization—the paper attempts to draw the line between what can be delegated to the machine and what must remain in human hands. Through this division of labor, the project aims to build, within its remaining term, a platform on which one can see through the text of the Tattvasaṅgrahapañjikā to the intellectual network behind it.