Written by Holly Smith, Collections Assistant for the Leeds Archive of Vernacular Culture within the University of Leeds Special Collections.

In working closely with catalogue descriptions during my role as collections assistant for the Leeds Archive or Vernacular Culture, I became aware of areas of sensitivity and problematic language in the collection. Whilst operating from home during lockdown 1.0 I began to research sensitive language in archive descriptions, which led to forming a workflow that has now been integrated into our approach as a University Special Collections.

Researching the Context

I found a lot of helpful articles and sources online and the social climate of Black Lives Matter and decolonisation meant a lot of interesting discussions were happening in the archive world.

It was also important to know my collection and the possible problem areas. With LAVC I knew that some folksongs included racist language, references to Gypsy Traveller and Roma communities often used historic labels, and that some of the farming practices could be seen as graphic content. This awareness honed in the process of researching, finding, and approaching these records.

Fieldworker's notebook containing drawings of old fashioned pig slaughtering tools.
We have a number of records based on agricultural practices that may be sensitive to modern audiences, such as these diagrams from a fieldworker’s notebook. “Skelton Response Book – 6Y3” (LAVC/SED/2/2/6/3/10) by Peter Wright. Image credit Leeds University Library.

Finding the Records

Searching for sensitive records is never straightforward and can often be unpleasant. It requires forming a reference list of offensive terminology most likely to appear in your collection and systematically searching description fields.

We have experimented with text analysis software as a more systematic way to search through records – a potential that we will hopefully unlock further with our upcoming National Archives Testbed Funded project.

Approaching Editing

We were aware that by discarding sensitive language we were removing the potential for second-hand trauma and misrepresentation, but by doing this we could also be interfering with the integrity of the record and covering up the mistakes of the past.

To avoid this, we felt it was important to differentiate between the creator’s voice and the curatorial voice. The creator’s voice reflects the original record and could impart important contextual information. In these cases, we decided that the language can be kept but placed in quotation marks to differentiate it from fact or current opinion. The curatorial voice can be influenced by biases from the archivist at the time. We viewed it as best in these instances to remove the language but ensure that you replace it with an equally meaningful, accurate and searchable term.

In all cases, sensitivity statements were attached and original descriptions saved as legacy records. These are available to be requested by researchers. This upholds the transparency of our sensitive language workflow and the integrity of our collections.