A corpus-based investigation of Italian Complex Nominals (CNs), of the form N+PP, which aims at clarifying their syntactic and semantic constitution, is presented. The main goal is to find out useful parameters for their representation in a computational lexicon. As a reference model we have taken an implementation of Pustejovsky's Generative Lexicon Theory (1995), the SIMPLE Italian Lexicon, and in particular the Extended Qualia Structure. Italian CN formation mainly exploits post-modification; of particular interest here are CNs of the kind N+PP since this syntactic pattern is highly productive in Italian and such CNs very often translate compound nouns of other languages. One of the major problems posed by CNs for interpretation is the retrieval or identification of the semantic relation linking their components, which is (at least partially) implicit on the surface. Studying a small sample, we observed some interesting facts that could be useful when setting up a larger experiment to identify semantic relations and/or automatically learn the syntactic peculiarities of given semantic paradigms. Finally, a set of representational features exploiting the results from our corpus is proposed.
Representing Italian Complex Nominals: A Pilot Study
Quochi V
2004
Abstract
A corpus-based investigation of Italian Complex Nominals (CNs), of the form N+PP, which aims at clarifying their syntactic and semantic constitution, is presented. The main goal is to find out useful parameters for their representation in a computational lexicon. As a reference model we have taken an implementation of Pustejovsky's Generative Lexicon Theory (1995), the SIMPLE Italian Lexicon, and in particular the Extended Qualia Structure. Italian CN formation mainly exploits post-modification; of particular interest here are CNs of the kind N+PP since this syntactic pattern is highly productive in Italian and such CNs very often translate compound nouns of other languages. One of the major problems posed by CNs for interpretation is the retrieval or identification of the semantic relation linking their components, which is (at least partially) implicit on the surface. Studying a small sample, we observed some interesting facts that could be useful when setting up a larger experiment to identify semantic relations and/or automatically learn the syntactic peculiarities of given semantic paradigms. Finally, a set of representational features exploiting the results from our corpus is proposed.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
