Analysing Academic Texts with Computational Linguistics Tools

Abstract

Writing academic texts in English introduces certain difficulties associated with translating Russian sentences with pronounced stylistic peculiarities, especially for young researchers who are just starting their publication activity. It seems impossible to study any genre without analysing examples of the discourse, which highlights the use of computational linguistics as it allows automating a lot of language and text processing mechanisms and generates relatively accurate quantitative results. The present study considers the application of AntConc and Coh-Metrix toolkits for analyzing master students’ abstracts to research papers written for international English-language journals or conference proceedings (Learner Corpus) in comparison with international researchers’ abstracts published in high-impact journals (Reference Corpus). The analysis conducted in the above-mentioned software tools revealed the drawbacks and strengths of master students’ texts, allowed characterizing them on the words, sentence and discourse levels, as well as outlined the potentials of their use in teaching academic writing skills.

Keywords

• academic writing
• scientific text
• computational linguistics
• automated text analysis
• antconc
• coh-metrix
• specialized corpus