The Tectogrammatical Representation of English in the Prague Czech-English Dependency Treebank Project
Speaker:Lucie Mladová
Institution: Charles University
Abstract:
The presentation aims to report on the tectogrammatical representation of English in the Prague Czech-English Dependency Treebank project (PCEDT 2.0). PCEDT 2.0 is a syntactically annotated parallel corpus of approximately 50 000 sentences adopted from the Penn Treebank (Wall Street Journal), translated into Czech. The annotation is based on the theoretical framework of the Praguian Functional Generative Description. The presentation focuses on our recent work in the English section, introducing the new tectogrammatical annotation manual and monitoring the statistics such as the numbers of sentences annotated and the interannotator agreement.