The Design of Croderiv 2.0

This paper deals withmethods applied in the expansion and design of CroDeriv – the Croatian derivational lexicon. The first version of the lexicon contained only verbs that were segmented and analyzed for morphemes. The database is available online. In a further development, other parts-of-speech (adjectives, nouns) are imported into the lexicon. All imported lexemes are analyzed in terms of their morphological structure and word-formation patterns. Due to new parts-of-speech, and a new type of information, the modification of the database structure was necessary. Here, we present a restructured version of the database, adapted to include other POS, and to explicitly mark word-formation patterns among derivationally related lexemes. We focus on underlying principles for precise and refined queries based on various parameters through the online search interface.