Lexical and grammatical semantics: a corpus-based statistical study of lexical semantic groups

The paper explores the correlation between lexical and grammatical semantics showcased by the inflexion of close meaning nouns of the Russian language. The basic assumption is as follows: the lexical semantics of a word determines the features of its morphological paradigm, and the totality of lexeme variations in use (the speech paradigm of the word) correlates with the unmotivated structuring of lexical semantic groups in a native speaker's mind. The correlation between lexical and grammatical semantics conveys itself through the concept of the word's grammatical profile, herein exemplified by cases of inflexion of several lexical groups of nouns. These are the words that refer to clothing, tools, as well as companies and organisations. The material for statistical and contextual analysis was retrieved from the National Corpus of the Russian language. Firstly, the statistical information is provided on the distribution of the case forms of nouns in removed homonymy zone of the National Corpus of the Russian language. Secondly, the regularities in the distribution of case forms of words are revealed that belong to the same lexical semantic group, and the ways this distribution is affected by the group's general semantics are found. The word groups concerned proved to differ in their grammatical profiles: the nouns of the "Clothes group" show peculiar statistical characteristics of the accusative and prepositional cases, the "Tools" group - in instrumental and accusative cases, the "Companies, organisations" group - in prepositional and accusative cases respectively. Thirdly, considering the distribution of case forms within each group, a "core group" of the words is identified where standard semantics is supposedly more prototypical. The analysis of grammatical differences found between the words of one semantic group, reveals certain aspects of compositional semantics, which are not always reflected in vocabulary definitions. Conclusions are made about the further possibilities of using case data for current studies of the correlation between lexical and grammatical semantics.