The project aims at an analysis of inter-sentential relations and discourse structure in Czech and is based on previous research work with the data from Prague Dependency Treebank (PDT). In the first steps, findings on Topic-Focus articulation in Czech will be tested in detail on the PDT data.In the second phase, the results obtained will serve as a basis for the study of coreference and inter-sentential semantic relations, and of the feasibility of the concept of salience applied on the analysis of discourse structure. This part of the research will result in a detailed classification of coreference types in Czech and in the enrichment of the annotation of the coreference in PDT.In the last step, the annotation of inter-sentential semantin relations in PDT and in Penn Discourse Treebank (University of Pennsylvania) will be compared and their compatibility will be evaluated.The linguistic studies will be complemented by formal approaches, using statistical as well as rule-based methods. The results of the linguistic research can serve as a base for an automatic processing of discourse relations in Czech. (en)
Analýza vybraných jevů aktuálního členění věty a koreferenčních vztahů s navazujícím výzkumem mezivětných vztahů a výstavby diskurzu. Analýza bude založena především na datech Pražského závislostního korpusu češtiny, ale i na dalším materiálu a na porovnání s anotačními systémy pro jiné jazyky.