Exploring Derivational Relations in Czech with the Deriv Tool

Authors

OSOLSOBĚ Klára HLAVÁČKOVÁ Dana PALA Karel ŠMERK Pavel

Year of publication 2009
Type Article in Proceedings
Conference NLP, Corpus Linguistics, Corpus Based Grammar Research
MU Faculty or unit

Faculty of Arts

Citation
Field Linguistics
Keywords NLP; word derivation; corpus linguistics
Description The aim of this paper is to present a tool for testing automatic word derivation for Czech. The derivation of some word formation types in Czech is in high degree regular. It can be described by formal rules. A new version of the web interface Deriv working with the morphological analyzer ajka enables us to formulate more complex word formation rules and to test more complicated cases of derivational relations. The second issue touched extensively in the paper are the types of derivational relations and their semantic classification. We have proposed 14 semantic classes for suffixes and 11 for prefixes. The tool Deriv helps considerably in establishing semantics of the derivational relations.
Related projects: