Structuring a specific domain: an electronic thesaurus of Ferdinand de Saussure's terminology

Although Ferdinand de Saussure's work has been disseminated all over the world, his thought was mainlyreconstructed and interpreted by his students and disciples. His authentic writings are however of paramount importance as they highlight the complexity of the philosophical and semiological system Saussure developed and the attention he payed to linguistic terminology. The great Genevan linguist himself often used a peculiar vocabulary. He forged neologisms or conferred a new meaning to existing words, used a few terms ephemerally, changed the denotation of some concepts over the years, so to create his own idiosyncratic terminology, which is now at the base of modern linguistics. In this paper, we present the first electronic thesaurus-lexicon of the Saussurean linguistic terminology that is being developed in the framework of the ongoing Italian project 'For a digital edition of Ferdinand de Saussure's manuscripts'. The lexicon population consists of an updated nomenclature, encompassing not only the terminology gathered by two famous Saussurean scholars, Robert Godel and Rudolf Engler, but also new terms extracted from recently found manuscripts under study. In the lexicon, the overall structure of Saussure's terminology is made explicit and the semantic import of its component terms as well as the nature and relevance of their relationships are defined. The lexical model, on which the thesaurus is grounded, is a customized version of the SIMPLE model, a de facto standard in the domain of Computational Lexicography that enables a highly structured representation of lexical knowledge. The first stage of the customization process consisted in designing and translating into the Web Ontology Language OWL the central component of the SIMPLE_FdS model, i.e. a domain-specific lexical ontology that would structure Saussure's terminology. The two other building blocks of the root model, i.e. a network of semantic relations and a set of semantic features, were adapted in order to meet the requirements of the domain of interest. The expressive means deemed suitable for our domain were maintained whereas additional specific features and relations were created to account for term properties and relationships peculiar to the conceptual organization of this domain of knowledge. So far, 375 terms were endowed with a rich semantic description. In the extended paper, examples of lexical entries will be provided and many interesting possibilities of lexical investigation will be pointed out. Information retrieval queries may in fact be formulated using any single piece of information encoded, be it a semantic relation, a feature or a lexical unit. Such an electronic thesaurus-lexicon, based on a multidimensional structuring of concepts and a large network of semantic relations among terms, is therefore, in our opinion, a most valuable lexical research tool. It is our deep conviction that it will contribute to better master some of Saussure's reflections and to gain a wider understanding of the overall domain terminology. It might therefore help shading light on original aspectsof the author's thought.