"...just as all rivers flow into the sea, all questions relating to the study of tales lead to the solution of the highly important and as yet unresolved problem of the similarity of tales throughout the world." -- V. Propp, in Morphology of the Folktale (1958)

The Geographic Lens on Stories (GLOS) project develops tools for studying how stories vary across cultures, places, and traditions. It combines long-standing geographic methods with emerging AI technologies to support structured comparison of narrative materials.

The project’s first phase produced machine-readable versions of the ATU and TMI folklore indexes and two pilot tools: Concept Matcher, which finds similar motifs or tale types for an uploaded text using embedding similarity; and ATU/TMI, which provides search and cross-reference across both canonical indexes.

The most recent phase—now paused—focused on creation myths, using LLMs to elicit and test conceptual schemas for cross-cultural comparison. Progress reports and experiment notes are available on the kgeographer blog, and the latest code is on GitHub.