"Conceptual maps and Computational Linguistics: the Italian ALTI project"

ALTI linguistic multifunctional databases are the result of a project started in 1998 when the research interests of different Italian universities (Università degli Studi di Napoli L’Orientale, Università di Pisa, Università degli Studi di Salerno, Università degli Studi di Roma “Tor Vergata”, Università degli Studi di Perugia) came together in one research project of national interest under the coordination of prof. Domenico Silvestri of the Università degli Studi di Napoli L’Orientale. ALTI stands for Atlanti Linguistici Tematici Informatici (Electronic Thematic Linguistic Atlases), which represent a new typology with respect not only to traditional lexicography, but also to computational linguistics, since they put together the characteristics of traditional dictionaries and terminological collections with a conceptual map, which highlights the conceptual relation among terms. The word atlas is used to underline that it is not a simple dictionary but a collection of maps, organized as multimedia hyper-textual atlases which collect linguistic data belonging to specialized linguistic areas (such as onomatology, food-terminology, numerals, linguistic activities, metalanguage of linguistics, lexical-grammar) in a multilingual and interlinguistic perspective. The Atlases describe the phenomenology of specific language areas by linking definitions and usage as given in conventional dictionaries to specific cognitive categories which create conceptual networks, several sets of maps (one for each atlas) or cognitive ellipses. These conceptual networks allow to navigate and to explore the relations between concepts inside a specialized linguistic area. On the one hand, they are meta-dictionaries, since they refer to other dictionaries, lexical and terminological resources already available in traditional or electronic format, but they also add new information by following original research perspectives, and, on the other hand, they offer multimedia information such as graphs, photos, films which represent very useful tools for the users. The languages investigated are: ancient and modern Indo-European and non-Indo-European languages; ancient and modern Celtic languages; Latin and Italy's ancient languages; major modern languages. In conclusion the Atlases are an open work since it is always possible to modify and update them with new contents and so achieve rich virtual cognitive universes. In our contribution we will describe the main features of the project, the research methodologies, the structure of the Atlases and of the lexical entries, the results achieved until now and the future aims. Index Terms: conceptual maps, Electronic Thematic Linguistic Atlases, computational linguistics