This paper argues that digital scholarly infrastructures, such as bibliographic databases and text corpora, are not neutral representations of reality but constructed models shaped by pragmatic, methodological choices. This is particularly critical in the "closing world" of proprietary platforms like Scopus and Web of Science, which present biased commercial products as objective scholarly tools. Through a critical discussion, this paper unpacks how unexamined decisions regarding operationalisation, standardisation, representativeness, and balance can distort our understanding of a scholarly field. In contrast, the paper highlights collaborative, open data ecosystems that leverage standardisation and interoperability. Ultimately, we argue that fostering a multiplicity of open, auditable models is essential to challenge the monolithic power of commercial actors, democratize research, and achieve genuine data sovereignty.
Maciej Maryl (Wed,) studied this question.