On the Most Frequent Sequences of Words in Russian Spoken Everyday Language (Bigrams and Trigrams): An Experience of Classification

Abstract

The article provides a description of the most frequent bigrams and trigrams obtained using the n-gram analysis technique on a representative sample of Russian spoken language. N-gram analysis allows identifying frequent lists of sequences consisting of n graphical words, which is important for describing corpus material in various theoretical and applied aspects. The source data for applying this technique was a sample of 388 episodes of everyday speech communication from the ORD corpus (about 110 hours of audio). The results of the n-gram analysis in the form of frequency lists of word sequences allow constructing a typology of the most common bigrams and trigrams in Russian oral communication and lead the study equally to the levels of grammar, pragmatics, lexicon, and phraseology. The list of the most frequent bigrams and trigrams contains grammatical structures (U TEBYA, YA NE PONIMAYU, MNE KAZHETSYA), idioms (in a broad sense of the term) (VSYO RAVNO, TO ZHE SAMOE), introductory units (TAK SKAZAT’, S DRUGOY STORONY), as well as a number of sequences typical only for oral speech, such as one-word pragmatic markers (NU VOT, KAK BY, NU V OBSHCEM), amplifications (DA-DA, TAK-TAK-TAK), and hesitations-vocalizations (E-E, M-M-M). The obtained frequency lists can be useful for solving many modern applied natural language processing tasks.