Non-Lexicalized Concepts in Wordnets: A Case Study of English and Hungarian

Here, we investigate non-lexicalized synsets found in the Hungarian wordnet, and compare them to the English one, in the context of wordnet building principles. We propose some strategies that may be used to overcome difficulties concerning non-lexicalized synsets in wordnets constructed using the expand method. It is shown that the merge model could also have been applied to Hungarian, and with the help of the above-mentioned strategies, a wordnet based on the expand model can be transformed into a wordnet similar to that constructed with the merge model.