The Atchan Song and Story Corpus: Towards a sustainable, adaptable online resource

Presented at the Language Documentation and Archiving Conference, Berlin & Online, 4-6 Sept, 2024 by Yao Maxime Dido, Siddharth Ganapathy, Lindsay Hatch, Rebecca Jarvis, Katherine Russell, Marie-Anne Xu.

This presentation introduces the Atchan Song and Story Corpus (Recueil de contes et chants Atchan/ÁCAN NANMƐ LÊ ÁLƐ? BHƆ? ), a community-oriented website presenting glossed and translated songs and narrative texts in Atchan (Kwa; spoken on the Ébrié Lagoon, Côte d’Ivoire). This project is a collaboration between researchers at UC Berkeley (USA) and University of Alassane Ouattara (Côte d’Ivoire), including one native speaker of Atchan. The website, currently under development, has two broad aims: to host recordings and transcriptions of songs and narrative texts, and to promote the recently-developed Atchan orthography. This presentation focuses on the workflow involved in managing and adding to the website. The primary data is hosted in two places: recordings are hosted on YouTube (with storytellers’/ singers’ consent), and transcriptions and translations are hosted in the online database Twisted Tongues (Ewert 2015). To support the generation and population of the website, we have written a Python script that converts text from TSV to JSON format, and a JavaScript program that generates the actual website. The script is adapted from the Moro Database script (Sande et al. 2016), with additional support for a bilingual interface (FR/EN), both IPA and orthographic representations, and video embedding. The website will be compiled via GitHub Actions, and the source code will be made publicly available. We expect the website to be live by summer 2024. The workflow is designed to support community involvement in the website, minimize technical knowledge needed to contribute to it, and promote longer-term sustainability: Twisted Tongues can be used online, and any appropriately-formatted TSV can be an input to the website. Because the source code will be public, the workflow could also be straightforwardly adapted to support use in other communities.