corpora

This article reports a corpus-based comparative analysis of the English-language transcripts of President Volodymyr Zelenskyy’s nightly video addresses during two one-month periods of the Russia–Ukraine war: 24 February–23 March 2022 and 24 February–23 March 2023. Two specialized corpora were compiled from the official presidential website (≈60,823 and ≈28,300 tokens, respectively) and processed in LancsBox.

We have studied statistical distributions of the frequency of sentences over their length for Ukrainian, Russian and English corpora and found the average sentence lengths in terms of linguistic signs, letters and words. It has been shown that the tails of the statistical distributions are satisfactorily described by the exponential function or the related ones, which is consistent with random nature of the sentence length.