From an analysis of 92,105,879 words in subtitles. Analysis was merely a count of lexical forms. For example, mü, mu, mı, mi (really all the same particle) are counted separately.