Terminologie et ontologie pour les humanités numériques : le cas des vêtements de la Grèce antique

This article presents the results of a collaboration between a classical philologist and a terminologist initiated during an EU-funded research project on ancient Greek clothing (Marie Skłodowska Curie Actions Project Chlamys. The Cultural Biography of a Garment in Hellenistic Egypt, grant 657898).The terminological ambiguity prevailing in the studies on garments of this period, whose sources are often incomplete (Delaporte 1981, 7), severely hinders the communication and knowledge sharing among experts. Clearly identifying what the terms designate, in what is specific, but also vague, as well as the naming of garments was at the heart of our research, which demanded the explicit representation of the knowledge that defines the different types of clothing.Added to this terminological concern is the need to enable the identification of a given garment according to its characteristics, i.e. to be able to identify (classify) the type or types of clothing to which it belongs. It should be noted that today there is no universally recognised (semi-)automated classification of garments as cultural artefacts of past cultures.This is why we are interested in the notion of ontology resulting from knowledge engineering and the semantic Web, and consequently in ontoterminology, a terminology whose conceptual system is a formal ontology. It was then necessary to define a methodology for the construction of such an ontoterminology, alternating the linguistic and conceptual dimensions. Ontoterminology defines concepts based on a set of essential characteristics, “knowledge primitives” shared by the experts, from which the ontology of the domain is built.The result is a multilingual terminology whose meaning is based on an ontology of the clothing domain of ancient Greece, which is seen here as shared and consensual, and which can give rise to many computer applications: multilingual terminology dictionaries, multilingual semantic search engines, automatic classification, document management, knowledge management, etc.This project highlighted the importance, even the necessity, of having a tool-assisted method based on a multidisciplinary approach, reconciling “emic” and “etic” approaches, and drawing on linguistics, terminology and knowledge modelling, a field of artificial intelligence. It also highlighted the importance of adopting a method that allows the collaboration of distributed end users, and whose results can be shared through the use of exchange formats complying with the standards in force, including those of the ISO and of the W3C.This article will focus on presenting the method and tool used in this project as they have generalisable features for similar projects in digital humanities, as well as some of the results obtained