Abstract
The paper deals with a well known problem of the distribution of grammatical markers within a certain category and whether this distribution is motivated semantically or not. It discusses the choice of singular and plural forms of nouns in Russian texts. Anna Wierzbicka was one of the first to recognize that the rules which regulate the usage of number markers in Russian are language-specific. She compared the Russian system with the systems of other languages, mainly, English and Polish (cf. Wierzbicka 1988). According to her theoretical assumptions, the difference of the linguistic behaviour of singular and plural forms is always semantically motivated, even if the corresponding nouns represent the same taxonomic class and occur in the same context (like different pieces of furniture, or kinds of vegetables, or types of dishes, etc.). The difference is always determined by the semantics of nouns and the typical functions of the extralinguistic objects they encode. This approach has been pursued by a number of scholars (see: Polivanova 1997; Lyashevskaya 2004а and others). Our paper develops this approach further on the basis of very different data. We take our data from texts written by nonstandard speakers of Russian (as defined by Mustajoki, Protassova & Vakhtin 2010) who are bilinguals. In our case these are students of the major universities in Almaty, who speak Russian fluently and acquired it long before entering university, having Kazakh as their native language. We have gathered and tagged a corpus of bilingual texts of this kind containing 60 000 tokens. This is a part of the Russian Learner Corpus (RLC) which facilitates identification of deviations from standard Russian. The analysis of these deviations is of particular interest for linguistic theory, since they provide cases of interference and interaction for languages of very different structure, such as Russian and Kazakh. The main goal of our paper is to show that the deviations in nominal number marking we come across in the texts of bilinguals are not arbitrary but motivated. They follow semantic strategies which are characteristic for speakers appealing to both systems at once. The paper argues that the violation of standard usage observed in the learner corpus can specify the rules governing Russian number usage which have been violated, even though the domain of number has been well described already.