This study investigated the mapping between Farsnet and Princeton English WordNet (PWN) and tried to answer these questions: Are the unmapped synsets in FarsNet with Princeton WordNet Persian specific lexical items? How can these cases be classified? Is it possible to find equivalents for some of these cases in PWN? In other words, were all of the equivalents of these lexical elements not lexicalized in English or there is another reason for that? For this purpose, a sample including 2000 synsets of FarsNet was selected and unmapped synsets were analyzed. Out of the sample, near 15 percent of the synsets had no mapping. The Categorization and analysis of the selected synsets showed that less than two third of these synsets were Persian specific words and more than one third  were synsets that were unmapped due to other reasons and therefore those are not specific lexical items in Persian. By enriching and completing the information in PWN and then FarsNet a lot of these cases can be mapped with PWN synsets.

Keywords: WordNet, FarsNet, lexical gap, lexicalization, mapping

Full-Text [PDF 1034 kb]   (2862 Downloads)