Publication at Faculty of Mathematics and Physics |
2010
Abstract
The research is based on a developing project that aims to annotate nominal coreference and bridging anaphora in the syntactically annotated corpus of Czech texts, PDT 2.0. In the process of annotating coreferential and bridging relations it became evident that the relatively low inter-annotator agreement is, to a large extent, due to the fact that a text may have a variety of legitimate objective interpretations, rather than the annotators’ mistakes or carelessness.
A classification of discrepancy types and possible causes of emergence thereof is presented. Typical examples of multiple interpretations of coreferencial and bridging relations in a text are given.