Non-standard numeral constructions in L2 Russian: A corpus-based study

Aims and Research Questions: The paper investigates variation in numeral constructions in the L2 Russian speech of bilinguals from different regions of Russia. The main research questions are the following: What factors prompt variation in this domain of grammar? Can we argue that non-standard marking is motivated by contact? Methodology: We conduct a corpus-based study of seven varieties of L2 Russian. The corpora contain spontaneous spoken texts collected between the 2000s and 2020s. Data and Analysis: Data from 181 participants born between 1920 and 1997, who are native speakers of 21 languages belonging to four different families, were analyzed based on a number of sociolinguistic and linguistic parameters that were modeled using hierarchical mixed effects logistic regression. Findings/Conclusions: We show that variation in numeral constructions is correlated with the speakers’ level of education and year of birth, the numeral type, and especially with the “collocationality” of numeral-noun combinations. The findings indicate that variation might be motivated by lower L2 proficiency and exposure to the input, but they do not provide strong evidence for contact influence. The negative correlation we found between higher “collocationality” and probability of non-standard marking supports a usage-based view of language acquisition as an experience-driven process in which frequency effects play a central role. Originality: This is the first study to provide a corpus-based analysis of this phenomenon in different varieties of L2 Russian. Implications: The findings have implications for the study of language variation in contact settings, and for the question of whether phenomena that apparently look like the result of pattern borrowing should rather be explained by other sociolinguistic and acquisitional factors that are not necessarily related to contact.