ERROR ANALYSIS IN MULTILINGUAL TRANSLATION PRACTICES

Abstract

Error analysis plays a critical role in improving multilingual translation quality between English and Russian. This study examines translation errors in both human and machine translation (MT) contexts (specifically Google Translate and DeepL), focusing on English–Russian and Russian–English directions. We adopt an IMRAD structured approach. In the Introduction, we highlight the significance of error analysis and present key theoretical frameworks such as contrastive analysis, error taxonomy, and interference theory. The Methods section outlines our comparative approach, combining contrastive linguistic analysis with an error taxonomy to classify errors (grammatical, lexical, semantic, pragmatic) in human translator outputs and MT outputs. In Results, we identify common error types in both EN→RU and RU→EN translation, supported by examples from published studies and corpus analyses. Typical errors include grammatical mismatches (e.g., articles, agreement, word order), lexical mistranslations (false friends, idioms), semantic inaccuracies, and pragmatic/contextual misrenderings. A comparison between human and MT practices reveals that while human translators are influenced by linguistic interference, MT systems often struggle with context and idiomatic usage. In the Discussion, we consider the practical implications of these findings for translators, educators, and MT developers. The study underscores the importance of error analysis for enhancing translation training and improving machine translation systems.