Internal authors: 
Issue Date:  2009
Abstract:  The high degree of variability in language use, especially in non literal utterances, represents a problem for formal semantics and for the attempts of simulation of the linguistic behavior in the field of Artificial Intelligence. Natural Language Processing (NLP) systems, in fact, have been until now heavily based on compositional mechanisms, that is mechanisms that assign meaning to complex constructions based both on the meaning of their components (meanings that are listed in a lexicon) and on restrictions on how the meaning of these components interact. Because of this choice -both conceptual and by design- metaphors and other forms of non literal meanings have been considered an inescapable problem and an obstacle by computational systems based on those principles (Carbonell 1982). In this paper, we describe a corpus-based method that gets along without deep compositional analysis and try to cope with the problem of variability in metaphors and the way in which they systematically appear in ordinary language and frames our thought.
URI:  http://hdl.handle.net/11584/34682
ISBN:  13689223
Type: 4.1 Contributo in Atti di convegno

Files in This Item:
There are no files associated with this item.

Questionnaire and social

Share on: