A corpus-based model of lexico-syntactic compensation in Russian and English academic discourse.

Abstract

Aim. To identify and model the systemic divergence of coherence strategies in English and Russian academic discourse within the Systemic Functional Linguistics framework, and to formalize the Principle of Lexico-Syntactic Compensation. Methodology. A balanced tripartite corpus of 1.3 million words was analyzed, consisting of original articles from high-impact journals (Scopus/WoS and VAK) and parallel professional translations. The study employed automated marker extraction with rigorous manual validation using Cohen’s kappa coefficient (к = 0.89), calculation of lexical diversity via Standardised Type-Token Ratio (STTR), and the Log-likelihood statistical test. Results. The analysis revealed that English discourse relies on a Syntactic Explicitness Strategy (high frequency of discourse markers), whereas Russian discourse favors a Lexical Compensation Strategy (high lexical variation). It was established that translators intuitively compensate for syntactic deficits using lexical means. Research implications. The concept of register-specific "rhetorical grammars" is substantiated. The practical significance lies in the model's applicability for evaluating and fine-tuning Natural Language Processing (NLP) systems and improving specialized translation quality.