Abstrakt
The manual describes how to annotate Czech learner texts using a multi-domain error tagset with a focus on morphology. The tagset was used to semi-automatically annotate a subset of texts from the CzeSL-man corpus.
See http://utkl.ff.cuni.cz/dokuwiki/doku.php?id=czesl:czesl for details about the CzeSL project.