From Identification to Remediation: Applying MaRMAT and Developing Lexicons for Reparative Description

Examining the metadata of a collection in an institutional repository of student work from a biological research center. Within the metadata, there are potentially harmful language and description items. It needs further investigation as part of a larger remediation process. This project used a reparative assessment tool to identify potentially harmful language and description items in metadata and to further develop a custom lexicon for remediation using reparative description. The MaRMAT tool [https://www.marmatproject.org] was used to create a custom lexicon list by identifying derogatory terms, offensive content, and descriptions from various governmental and non-profit sources, as well as pre-curated lexicon lists. The concept of reparative description was used to create terms related to Indigenous communities. Based on the results, 960 occurrences of problematic content were identified within the collection's metadata, including those that unintentionally spread Indigenous communities’ Traditional Knowledge, locations of endangered species, and other sensitive information. The custom lexicon list will enable nuanced discussion of content censorship, historical context, and further engagement with the CARE principles. Additionally, the list will help with the contextual and data analysis of the collection. Next steps will be discussed for the remediation process.