Constructing a learner-friendly corpus-based dictionary of Serbian verbal aspect

This paper describes ongoing work on the construction of a dictionary of Serbian verbal aspect based on data from corpora and specially adapted for the needs of L2 learners, who face the difficult task of choosing the appropriate aspect and the correct morhological form of the verb. We explain the process of automatic extraction of information about aspectua behaviour of verbs from a corpus. We show that a quantitative notion of productivity of verb forms based on the extracted corpus data id useful for a systematic and rich representation of the knowledge about Serbian verb aspect