The Text Frequency of Czech Noun Declension Patterns

Abstract

It is generally accepted that the frequency of words (as elements of a lexicon) in texts follows Zipf’s Law. Our previous research has shown that Zipf’s Law also holds quite well for the morphological pattern distribution over the lexicon. An interesting question arises: If the text frequency is relatively high for words following morphological patterns with low lexicon frequency, then what does the text frequency distribution of the patterns look like? We have made an experiment to get some idea of this distribution by combining a morphological dictionary of Czech words with the data available in the Frequency Dictionary of Czech, containing 50,000 of the most frequent Czech words.