Abstract

In this paper, we challenge the vision of creating a giant knowledge graph of reusable research data from multiple projects and platforms, based on Semantic Web technologies and a shared conceptualisation in the form of a commonly used ontology ecosystem, promoted by the Semantic Data for Humanities and Social Sciences project (SDHSS). Three main issues are raised: the difficulty of proposing a cross-disciplinary ontology for humanities and social sciences given the different research agendas and constructivist approaches; the difficulty of interlinking the objects present in the different information systems, due to different notions of identity; the challenge of dealing with uncertainty, inconsistencies and contradictions in information sources, when the aim is to produce ‘factual’ information. First, I propose an epistemological analysis of information, as distinct from data and knowledge, then I develop a conceptualisation of information in humanities and social sciences, based on the OntoClean methodology, and the foundational ontologies DOLCE and Descriptions and Situations, and show how it can ground a cross-disciplinary ontology ecosystem based on the CIDOC CRM and its SDHSS integration. Finally, I apply these principles to a detailed analysis of the information production process and show how the proposed conceptualisation provides possible solutions to the three challenges raised.