Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

Learning From, Talking About, and Reflecting on Data Loss: A failure report

Domaine:

natural language processing

Type de record:

dataset
Créateur:
Harvey, Andrew
Éditeur:
Zenodo
Hôte:avatar
Griscom and Harvey (2020) is a large, diverse, archival record of language and cultural material from the Hadza people of Tanzania. Collected between 2019 and 2021 and employing a community-based model, while this collection represents a considerable contribution to Hadza language work (c.f. Leonard 2021), it is not perfect: during the course of collecting and processing the audiovisual materials, a considerable amount of video material associated with the collection was lost. This loss of data will be the focus of the talk. Language documentation projects are complex undertakings (c.f. Bowern 2008: 48), and failures – including data loss (c.f. Tsutsui-Billins 2019) – are frequent. At the same time, it is not always easy for language documentarians to talk about their failures – this ultimately makes it easier for specific failures to occur again, at different times and in different projects. Modelling this talk on the failure report (c.f. Engineers Without Borders Canada 2017), the specific failure is examined in detail, the causes of the failure made explicit, and the lessons learned enumerated. The project which resulted in Griscom and Harvey (2020) is, in many ways, a great success, both in terms of data collected, as well as the relationships developed throughout. With that said, focusing on the data loss which occurred is useful in several ways: it helps make what happened explicit such that learning can occur, it creates a space to talk about the dynamics of community-based work with marginalised communities, and it allows us to be reflexive in our praxis. Note: This talk has not gone through a process of peer review, and findings should therefore be treated as preliminary and subject to change. Acknowledgement and citation: Harvey, Andrew. 2022. Learning From, Talking About, and Reflecting on Data Loss: A failure report. Talk given at the Language Documentation and Archiving Conference, Berlin-Brandenburg Academy of Sciences, Berlin, Germany. 07/10/2022.

Visit

doi.orgzenodo.org

Languages

Hadza

Tags

Languages of TanzaniaTanzanian Rift ValleyLanguage Documentation and DescriptionHadzaData LossFailure Report

Licenses

Creative Commons Attribution 4.0 Internationalhttps://creativecommons.org/licenses/by/4.0/legalcodeOpen Accessinfo:eu-repo/semantics/openAccess