Annotating foreign learners’ Czech

One of the challenges of contemporary corpus linguistics is the compilation and annotation of corpora consisting of texts produced by non-native speakers. In addition to morphosyntactic tagging and lemmatisation, such texts can be annotated by information relevant to the specific nonstandard use. Cases of deviant language use can be corrected and identified by a tag specifying the type of the error. Because of the properties of Czech, namely rich inflection, derivation, agreement, and a largely information-structure-driven constituent order, it is not straightforward to design an annotation scheme satisfying all requirements on the description of errors produced by non-native learners. Our proposal aims at an optimal solution that is still realistic given the annotation costs and the demands of the corpus users. After an overview of issues related to learner corpora in §2 and a brief introduction to the project of a learner corpus of Czech in §3 we present the issues of annotation in §4 and the concept of our annotation scheme in §5, followed by a description of the annotation process in §6.