We study rank-frequency relations for phonemes in texts written by different authors. We show that they can be described by generating phonemes via random probabilities governed by the (one-parameter) Dirichlet density, the simplest density for random probabilities. This description allows us to demonstrate that the rank-frequency relations for phonemes of a text do depend on the author. The author-dependency effect is not caused by common words used in different texts. This suggests that it is directly related to phonemes or/and syllables. These features contrast to rank-frequency relations for words, which are both author and text independent and are governed by the Zipf's law.
View on arXiv