Annotation of Analytic Verb Forms in Czech – Complex Cases

Abstract The article deals with complex cases of determining the attribute verbtag, which contains the values of morphosyntactic categories of analytic verb forms. The latest corpora of contemporary written Czech from the SYN series are tagged with this attribute. In this paper, we focus on cases where it is difficult to identify values of verbtag categories. These include, e.g. the identification of the auxiliary verb být ‘to be’, recognition of the mood and tense of coordinated participles, or determining the number in compound forms in which the individual parts have a different morphological number. Some of the problems are of a theoretical nature, since it is not clear what the correct solution should be. Here we have arbitrarily opted for one option that was offered. Other problems are due to imperfections in the algorithms we use for annotation. The solution here is to improve these algorithms.