HIGH-FREQUENCY VERB-NOUN COLLOCATIONS FROM DICCTIONARIES IN THE RUSSIAN NATIONAL CORPUS

Corpora and statistical tools have provided new opportunities for further studying the collocability of lexical units. In this context, stable word combinations are described in detail in explanatory and specialized dictionaries. This paper analyzes how, on the one hand, dictionary collocations (units fi xed in lexicographical sources) are refl ected in the corpus (based on the Russian National Corpus) and, on the other hand, how frequentword combinations retrieved from the corpus correspond to dictionary data. The material includes a list of collocations constructed on the “verb + noun” model, selected from a number of dictionaries of the Russian language, with the following nouns: zhizn’ ‘life’, sila ‘power’, delo ‘business’, slovo ‘word’, rabota ‘work’, vremja ‘time’, vzglyad ‘glance’, vorpos ‘question’, vozmozhnost’ ‘opportunity’, pravo ‘right’. The ranking of word combinations in the corpus was done using the frequency of co-occurrence, as well as logDice. The results show that high-frequency units from the corpus are quite well represented among the collocations presented in dictionaries (about 67%). Conversely, more than half of the identifi ed frequent word combinations (55%) are dictionary collocations. The logDice measure demonstrates results similar to ranking by co-occurrence frequency.