This paper discusses a recent project which applied computational methods to catalogue data in an attempt to generate new information on the provenance of Islamicate manuscripts in UK repositories. Using a subset of records taken from the Fihrist Union Catalogue as a dataset, we analysed and grouped together manuscript descriptions which shared selected physical features, then examined the occurrence of records with secure provenance within those groups to see if information on the place of origin could be extrapolated across them. While we gained useful information regarding the provenance of manuscripts, the chief conclusion of the project was that catalogue data, in its current state, poses serious challenges for quantitative analysis. This is partly due to the various purposes for which data has traditionally been collected (in contexts where codicological descriptions had a different purpose) and partly due to inconsistencies in the dataset. In our conclusion we put forward strategies for working with inconsistent data, make suggestions for changes to cataloguing practices to answer the requirements of digital methods, and propose new research questions addressing the history of catalogues and cataloguing practices which came into focus during the project. We also make a case for the potential of digital methods to enable new approaches to decolonisation, focusing on data modelling, data provenance, and accessibility.
Jones et al. (Fri,) studied this question.