A Morphological and Syntactic Wide-coverage Lexicon for Spanish: The Leffe

International Conference RANLP 2009 - Borovets, Bulgaria, pages 264–269 A morphological and syntactic wide-coverage lexicon for Spanish: The Leffe Miguel A. Molinero Grupo LYS Univ. of A Coru˜na A Coru˜na, Spain mmolinero@udc.es Benoˆıt Sagot Project ALPAGE INRIA Paris, France benoit.sagot@inria.fr Lionel Nicolas ´Equipe RL Laboratoire I3S Sophia Antipolis, France lnicolas@i3s.unice.fr Abstract In this paper, we introduce the L´exico de Formas Flexionadas del Espa˜nol (Leffe), a wide-coverage morphological and syntactic Spanish lexicon based on the Alexina lexical framework. We explain how the Leffe has been created by merging together several heterogeneous lexicons and how the Alexina lexical framework has been applied to Spanish. We also introduce a semi-automatic technique based on a tagger to detect the lexicon’s deficiencies. preliminary evaluation shows the potential of the LeffeA and the relevance of both creation and extension processes. 1 Introduction High-level Natural Language Processing (NLP) tools require reliable linguistic resources, such as lexicons and grammars. Nowadays, such relevant resources exist for English, but are often absent or incomplete for other languages, even major ones. For example, some lexical resources exist for Spanish, but none of them combines satisfactorily the following properties: • coverage: all words, including rare ones, in all categories should be included; • quality: manually and automatically developed resources contain various errors; • richness: applications such as (deep) parsing require at least morphological and syntactic information, including subcategorization frames. The Leffe1 is a wide-coverage morphological and syntactic lexicon based on the Alexina framework [13, 15, 1]. This lexicon follows the linguistic criteria applied on the French lexicon Lefff2 taking advantage of the linguistic proximity between Spanish and French as Romance languages. The main contributions of this piece of research are the following: • we present a morphological and syntactic wide coverage lexicon for Spanish; • we describe an enhanced available lexical framework, 1 L´exico de formas flexionadas del espa˜nol - Lexicon of Spanish inflected forms 2 Lexique des formes fl´echies du franc¸ais - Lexicon of French inflected forms • we expose a simple semi-automatic PoS tagger-based approach to detect numerous missing entries in a lexicon (including homonyms). The work described here is one of the starting points of the recently created Victoria project. This project aims at developing techniques and tools for an efficient acquisition and correction of the linguistic resources necessary to symbolic syntactic parsers. The first phase of the project focuses on Spanish, Galician3 and French. This paper is organized as follows: we present first a brief description of Spanish language in Section 2. In Section 3 we describe the lexical framework used to formalize the linguistic information. Then, we briefly describe how some available resources were used to develop the Leffe. In Section 5 we present a semi-automatic technique to correct and extend the lexicon. Finally in Section 6 we show preliminary evaluations of the lexicon and present our conclusions in Section 7. 2 The Spanish language in brief Spanish is a Romance Language, just like Italian, French, Portuguese, and many others. Despite being spoken as a mother tongue …

Content truncated to 500 words