A method of semantic change detection using diachronic corpora data

Abstract

© Springer Nature Switzerland AG 2020. The article proposes a method for detecting semantic change using diachronic corpora data. The method is based on the distributional hypothesis. The analysis is performed using frequencies of syntactic bigrams from the English and Russian sub-corpora of Google Books Ngram. To obtain the word co-occurrence profile in its new meaning, syntactic bigrams that contributed most to the word distribution change are selected and their time series are clustered. The method is tested on a group of English and Russian words which gained new meanings in the 20th century. The obtained results show that the proposed method allows one to detect semantics changes, as well as to determine the time of these changes.