Från kränkande till inkluderande : En dokumentstudie av metadataprojektet DE-BIAS

Sammanfattning

Archives are never neutral repositories of a final and correct version of history. Their systems, metadata and policies are made in a specific social and political context. Archives are powerful institutions where what is saved and how it is catalogued and organised impacts what type of research is possible and what types of conclusions a researcher can draw from its material. Marginalised groups such as women, people of colour, queer people and others that do not fit the norm are often either absent from the archives, or the metadata describing what is there is inadequate, offensive, biased or incorrect. As a part of a democratic society, archives need to reflect and represent diverse people and diverse stories. The purpose of this thesis is to explore how metadata can be contextualised and complemented in cultural heritage databases to make it more respectful, nuanced and correct. This study will be based on the pan European metadata project DE-BIAS. The research questions focus on how contemporary critical perspectives can be used in archival metadata so as not to reproduce harmful and offensive norms, and how communitybased development can contribute to this end. This case study uses document analysis methodology to represent and interpret the official documentation that the DE-BIAS project has produced and published on their webpage. An in-depth analysis and discussion of the material will be made with the assistance of critical archival theorists like Michelle Caswell, Joan M. Schwartz and Terry Cook. The theoretical framework will center archives as power, power imbalance and ways to use critical theory to change both praxis and possibility for change. This thesis concludes with recommendations for best practices on how to approach bias in metadata in a Swedish context. First and foremost, the process needs to start from within the institution, with space provided for staff education in critical and intersectional thinking around bias. Secondly, a specific Swedish vocabulary needs to be developed as which words are seen as offensive often depends on cultural context. Using this vocabulary in conjunction with an AI-tool like the one developed by the DEBIAS project will make it financially possible to augment bias metadata. The last section is focused on community and how continually working together with marginalised groups is the key component to making a functional and relevant vocabulary, educational workshops and future projects that aim to mitigate bias.