This paper describes the vocabularies used in PARTICIPATION, a Horizon2020-funded project aimed at preventing extremism, radicalization, and polarization. To fully take advantage of Linked Data, all data in the project need to be expressed in a semantic format, and all annotation services should be accessible through a semantic API. Most of the data can be expressed by extensively leveraging common vocabularies in the Linguistic Linked Data sphere. However, certain key concepts were not present in any of the popular vocabularies, such as ideologies, morality, and narratives.
Some types of analysis also required the use of resources aligned with Linguistic Inquiry and Word Count (LIWC) software. As a result, four vocabularies were developed: Senpy Annotations, SLIWC, Morality, and NarrOnt.
Senpy Annotations is a vocabulary designed to represent any kind of annotation in the context of NLP services and resources. SLIWC is a vocabulary and SKOS taxonomy that aims to represent LIWC dimensions. The NarrOnt (Narrative Ontology) vocabulary models the concepts of a narrative and an ideology linked to a piece of content. Lastly, morality is a vocabulary for expressing annotations that follow the Moral Foundation Theory (MFT). These vocabularies have been designed and published using Linked Data principles and best practices.
Most importantly, they follow an orthogonal design, integrate well with existing vocabularies, and only describe specific parts of a domain. We believe that the usefulness of these vocabularies will extend beyond the scope of this specific project.