摘要：The paper presents the Croatian WordNet, a computer lexicon of the Croatian language developed along the principles of the Princeton WordNet. In this paper we focuss on verbs and describe the principles we use in the construction of verbal synsets and lexical hierarchies based on hyponymy / hyperonymy. The paper also presents other semantic relations such as antoynymy, cause and subevent that we use in semantic fields of verbs. The paper also discusses the processing of so called aspectual pairs of Croatian verbs and offers a solution in terms of distinguishing temporal and semantic aspectual distinction. This has furher implication on the construction of synsets and lexical hierarchies. The analysis is demonstrated on the verbs of eating.
