Critical data provenance as a methodology for studying how language conceals data ethics

This paper contributes to a project that maps the concept of ‘data provenance’ into qualitative data research. Data provenance is a concept from computer science that allows us to trace the history of sets of data to determine the accuracy and validity of a database. We transform this into a critical concept by arguing that data provenance can be used to trace the acts of governance and rhetoric. We then analyse the forms of discursive power that shape transactions in data. Our approach can create a history of the governance and justifications that are used to build and assemble datasets from multiple sources. Presently, data often lacks this information about its own discursive origins, unlike other forms of data provenance. Critical data provenance gives us a model for thinking about how governance could be mapped onto data. And thus, critical data provenance is also a framework for critiquing the justifications used when dataset owners acquire data–for instance, whether data stewards or users are encouraged to be ‘open’, to ‘share’, or be ‘transparent’. To this end, we demonstrate this model of thinking and provide the analytical tools necessary to redeploy these ideas into new contexts.