Integrating Two Semantic Lexicons, SIMPLE and ItalWordNet: What Can We Gain?


Adriana Roventini (Istituto di Linguistica Computazionale del CNR - Pisa - Italy)

Marisa Ulivieri (Istituto di Linguistica Computazionale del CNR - Pisa - Italy)

Nicoletta Calzolari (Istituto di Linguistica Computazionale del CNR - Pisa - Italy)


WO17: Semantic Lexicons


In the last years, at the Institute for Computational Linguistics in Pisa, a few lexical resources have been developed aiming at encoding complex lexical semantic information. ItalWordNet and SIMPLE are two of these resources which, tackling semantics in the lexicon from different points of view, and being at least partially complementary, could certainly profit from linking each other. These resources in fact evidence different aspects of the lexical information: in SIMPLE, which adds a semantic layer to the morphological and syntactic ones developed in PAROLE, the connections between semantics and syntax are preeminent; ItalWordNet (as the Princeton WordNet and then EuroWordNet) is built around the basic notion of a synset and various semantic relations are encoded between synsets while syntactic aspects are not taken into consideration. In the paper we describe an experiment we carried out, aimed at exploring the feasibility of linking these lexical resources, being convinced that a noteworthy gain could be achieved through this operation. As we will show in the following some problems came in the foreground but also considerable advantages concerning the coherence and the completeness of both of them.


Lexical resources, Semantic lexicons, Integration

Full Paper