Publication at Faculty of Mathematics and Physics |
2019
Abstract
There exist several morphological dictionaries for Czech. They differ only in solutions of complicated morphological features.
Various attempts have been made to unify their approaches, but only some of them were implemented. The paper deals with several such features and compares their solutions taken in two different projects, namely preparation of the new edition of PDT (Prague Dependency Treebank) and NovaMorf.
The features presented in this paper are: aggregates (the word-forms without a clear part of speech, e.g. užs, oč, naň), and variants - inflectional (more wordforms for a particular combination of lemma and morphological tag) as well as global ones (mainly orthographic variants expressed in all wordforms of a paradigm).