Institute of Formal and Applied Linguistics
Charles University, Czech Republic
Faculty of Mathematics and Physics
HVar
Disagreement in corpus annotation and variation of human understanding of text
Disagreement in human annotation of complicated textual phenomena is ubiquitous but in many
cases, it is considered undesirable. In our project, we turn the disagreement into advantage and
understand it as an indicator of possible ambiguity of the text itself. This hypothesis is
tested by a systematic analysis of different understanding of particular textual phenomena, such
as discourse relations, coreference, and information structure. The analysis is carried out
on double / multiple corpus annotations covering different text registers in languages such as
Czech, English, and German. In this way, our project aims at a systematic comparative analysis
of differences in understanding of textual phenomena. We believe that the findings of the
project will contribute to achieve new insights into language structure in relation to
psycholinguistics and cognitive science.