Platforms as Epistemic Devices: The Production and Contestation of Knowledge in Platform Societies

Social research now depends on environments that commercial platforms own and continually redesign. This article argues that platforms operate not only as objects of inquiry and infrastructures for data production but also as epistemic devices that shape what can be observed, classified, measured, and anticipated in social research. Drawing on science and technology studies, digital sociology, and critical data studies, it identifies four operations through which this shaping occurs (selection, classification, quantification, and temporalisation) and integrates them into a five-level framework that traces how social activity becomes scientific knowledge. The account differs from existing treatments of method as co-production by specifying what makes the platform a historically distinctive device: it is proprietary, strategically self-interested, and able to withdraw the visibility on which research depends and to reorganise itself once it has been studied. The closure of the research API provides the clearest expression of this condition, while Article 40 of the Digital Services Act marks an emerging attempt to subject epistemic access to public regulation. Offered as a reflexive instrument for research design rather than a procedure, the framework grounds a case for methodological pluralism and is set out together with its limits and possible empirical extensions.