Abstract
In this paper, we describe a corpus based experiment focused on the possibility of identifying Czech light verbs. The experiment had two main aims: (i) to establish the inventory of Czech light verbs entering into combinations with predicative nouns, and (ii) to verify the adopted criteria for distinguishing light usages from full usages of the given verbs.
As for the inventory of light verbs, we propose and verify the hypothesis that the possibility of light usages of verbs is related to their semantic class membership rather than to their high frequency. In the second part of the experiment, we exploited the compiled inventory of Czech verbs inclined to occur as light verbs.
The criteria adopted for distinguishing light usages of a verb from its full ones were applied to selected corpus sentences with these verbs. Three annotators were asked to determine whether a verb occurrence in an extracted corpus sentence corresponds to the full or to the light usage of the given verb.
The feasibility of this