Abstract
The interest in the use of corpora in teaching Czech as a foreign language has gained strength in the last decade, however, no comprehensive study of vowel length based on corpus data has appeared yet. Therefore, in this paper I conduct the first attempt to capture the most common difficulties foreigners have with the vowel length on the basis of corpus material.
I present one example for each of the three autosemantics: nouns, verbs, and adjectives. Within the nouns I discuss a problem with the vowel length in the locative plural.
For the verbs I focus on the vowel length in the past participle. In the case of adjectives, I pay attention to lexemes ending on -ální, -ární a -étní.
Moreover, I describe the procedure of data analysis and suggest concrete exercises for students that can be used in the classroom. Corpus teaching can be accessed directly and indirectly, therefore we offer both options within this paper.
It should be emphasized that corpus approach in teaching can only be a complement to traditional teaching or can be implemented in addition to traditional teaching in seminars fully focused on data-driven learning.