Towards an analysis of the epistemic frameworks of big data

In this paper we explore some epistemological discussions about big data. As our main epistemological concept, we use the notion of Epistemic Framework formulated by Piaget and Garcia, and reviewed by us in more recent works, which seeks to problematize the conditioning of ethical and political values, as well as other regulations, in the development of scientific knowledge. We propose to order the debates in the recent literature on big data through various dualities, such as data/interpretation, automatic/non-automatic, or causality/correlation, around which it is possible to view two specific epistemic frameworks, tending to split these terms, or to their integration. On this last mode of focusing the debates, we identify some epistemological challenges of research with big data, especially in the computational social sciences.