Date of Degree

9-2021

Document Type

Thesis

Degree Name

M.A.

Program

Linguistics

Advisor

Kyle Gorman

Subject Categories

Computational Linguistics | Russian Linguistics

Keywords

Russian loanwords, inflectional morphology, loanword detection, finite-state transducer, language model

Abstract

This paper investigates recent English loanwords in Russian and explores ways in which computational methods can help further theoretical research. The goal of the study is two-fold: to find new, previously unattested loanwords borrowed over the last decade and to examine the rate of adaptation of the new borrowings, attested by the degree to which they conform to the constraints of the Russian language. First, we train a finite-state pipeline that combines character n-gram language models, which encode phonotactic and lexical properties of loanwords, with a binary classifier to detect loanwords. The model achieves state-of-the-art performance results during evaluation, surpassing previously established benchmarks. Secondly, we introduce a new and extended corpus of recent Russian loanwords that have been detected in Web texts by our model. The corpus includes loanwords together with their morphological features, part-of-speech tags, and sentences in which they occur. We conduct an analysis of inflectional morphology of the identified loanwords, investigating the rate of indeclinability of recent loanwords and stem-final consonant alternations in verbs.

Share

COinS