NovaMorf: the end of a long period of convergence and divergence in the processing of Czech morphology
Abstract
In this text, we want to outline the coincidences and differences of the two tagsets used for automatic morphological analysis of Czech. We will show how much the originally unintentional and time-sustained double-knee of the so-called Prague and so-called Brno systems can be overcome in the foreseeable future under the NovaMorf project.
We will look at the relationships between the branding of morphological categories and values in the NovaMorf proposal compared to the two older systems. We base our assessment of the Brno system on an article by the Czech Morphological Tagset Revisited. (Jakubíček, Kovář, Šmerk, 2011).
We base our knowledge of the Prague system on a description of Prague's position tagset (see http://ufal.mff.cuni.cz/pdt/Morphology_and_Tagging/Doc/hmptagqr.html) and on a monograph by Jan Hajic (Hajic, 1994, 2004). Our goal will be to show how the experience of using both systems has resulted in an effort to inspire positives and avoid failed solutions on both sides (Srv.
Osolsoba et