Following the footsteps of synthetic data: documenting epistemological justifications in paradata

Data has both ontological and epistemological justifications. While the ontological ones are intuitively crucial, the importance and especially consequences of the variety of epistemological ones might remain less obvious. Especially with synthetic data, its becoming might appear as a much more straightforward process in relation to the ontological uncertainty of to what extent it is capable of standing for ’real’ data. Still, data is as it is because of how it has come into being. The becoming is infrastructural to the data comparably to how its infrastructures are critical to its becoming, and how understanding the becoming is infrastructural to its informed use and its infrastructures are critical to making synthetic data to work as intended.

The presentation delves into the question of how to document and convey knowledge – in terms of paradata – about the making, processing and earlier use of synthetic data, how paradata is linked to the contexts, practices and infrastructures of its becoming, and how the multiple contextualities of paradata affect the epistemological justifications it can provide to synthetic data. The presentation draws from an inquiry into a selection of synthetic data sets in the domain archaeology to show how synthetic data can be epistemically highly fragile, rely on the functioning of an intricate set of in multiple senses invisible and only momentarily stable infrastructures, and how its fragilities can be difficult to make visible enough to be communicated to the stakeholders of the data and whatever comes out of its use.