[4I4-OS-17a] OS-17

2:15 PM - 2:30 PM JST(5:15 AM - 5:30 AM UTC)

[4I4-OS-17a-04] Synchronic and Diachronic Exploration of the Classical World Centered on CitiesAI-Collaborative Support for Textual Interpretation Based on People, Time, and Place

〇Ikko Tanaka¹, Jun Ogawa², Naoya Iwata^3,4 (1. J. F. Oberlin University, 2. Univ. of Tokyo, 3. Nagoya University, 4. National Institute of Informatics)

Keywords:

Western Classics,Large Language Models,Geospatial data

This study reconceptualizes cities as nodes of intellectual activity in Western Classical studies by constructing an environment that integrates and visualizes data on authors, chronology, and geography. Authorial metadata are derived from existing textual resources and knowledge bases using a data construction approach that combines large language model–based extraction and candidate suggestion with human verification. Temporal information, including birth and death dates or periods of activity, is represented as estimated time ranges rather than fixed values, enabling analyses that explicitly account for uncertainty inherent in historical sources. Geographical data are linked to Pleiades identifiers, while city and place-name variants are manually reviewed in accordance with historical evidence and scholarly conventions. By synchronizing timelines of authorial activity with maps, the system enables intuitive identification of which individuals were associated with which cities at particular times, supporting synchronic and diachronic comparisons centered on cities. In addition, retrieval-augmented generation is applied to Western Classical databases for selected cities and periods to present interpretations grounded in relevant textual sources, enabling the reconstruction of intellectual history from cities as analytical anchors.