Towards an Efficient Representation of Restricted Lexical Cooccurrence

In this paper, we address the problem of efficient representation of restricted lexical cooccurrence information in the lexicon. In a study carried out with forty German emotion lexemes, significant correlations between the semantics of these lexemes and their verbal collocates have been found. Drawing on these correlations, we suggest a more efficient representation of collocation information in the lexicon: recurrent collocates should be extracted from the individual lexical entries and listed together with their semantic constraints in the entry for what we call the generic lexeme of the semantic field under study.