Introductory course I: Linked Data and Ontologies
Post by: Mariana Velázquez
Classical Western philosophy developed the concept of “ontology” as a means of organizing knowledge by studying the categories of being and the relationships among those categories. Because of its highly theoretical nature, the study of ontology had been strongly attached to the philosophical branch of metaphysics. Among the principal questions that ontology addresses, we find what is a thing, how is it, how much it is, and where it is in relation to other things. Today, the classical concept of ontology has not aged, but rather, it has been appropriated by the field of information and computer science.
The first part of Victoria Eyharabide’s introductory course, Linked Data and Ontologies, revolved around the description and explanation of the role of ontology in the metadata world. Her course touched upon the discussion of the main components of ontologies (individuals, classes, and properties) in order to improve Musiconis’s search engines. To some extent, the exercise reintroduced the main philosophical questions of ontology within the context of the Musiconis database. In other words, by expanding the description of the images contained in this database and establishing new relationships among them, the images will become more searchable and accessible to researchers.
The working platform is Protégé 5.1.0 (developed by Stanford University). We were provided with the basic skills to understand the program and its possibilities. The first thing we were taught is that natural language is ambiguous and for that reason, we had to work with both semantics and computer language (specific codes and categories). The result is a kind of language that is supported by reasoners and at the same time, it expresses a variety of concepts and restrictions.
Such language is based on the OWL (Web Ontology Language), a family of knowledge representation languages that enables the creation of ontologies. We focused on the variant OWL DL to conform several syntaxes when describing the images. The Protégé tutorial encompassed 17 exercises that revolved around practicing different ways of describing a class, an entity, a concept, and an individual. By establishing a hierarchy of these components, we were successfully able to comprehend and visualize the ontology of a specifically given concept and to model a structure of knowledge around it.
Eyharabide’s introductory course emphasized two main aspects. On the one hand, the course entailed classical ontological questions about what constitutes a specific object or image, while on the other, it involved the need of transferring those questions to the realm of computer science and information. In short, we learned how to teach a computer to understand and apply the complexity of an ontology. Certainly, this is one of the pivotal challenges faced by the growing field of Digital Humanities which has transformed the ways in which we study and approach medieval cultural production.