Author’s name:
Olga V. Petrova, Maria Yu. Rodionova – Linguistics University of Nizhny Novgorod, Nizhny Novgorod, Russia
Abstract:
This paper addresses the problem of reproducing such fundamental textual properties as coherence and cohesion in machine translation. The research material consists of texts translated by three different computer translation programs, which are analyzed in order to identify the causes and consequences of a machine’s inability to adequately reproduce these text-forming categories. The study demonstrates that the principal reason for violations of textuality lies in the machine’s inability to perceive a text as a unified structural, semantic, and communicative whole whose components are interconnected by numerous intra-textual links. An additional factor is the machine’s inability to comprehend context in its full complexity. For most computer programs, context is often reduced to the subject domain of the text, while coherence and cohesion may, in some cases, be maintained only at the sentence level. In other instances, these categories are violated even within a single paragraph. As a result of the machine’s failure to identify intra-textual links, such links are frequently reproduced incorrectly in translation. Erroneously established or inadequately rendered intra-textual connections may lead to distortions of content and meaning, which are sometimes concealed behind an apparently correct grammatical or lexical form. Moreover, the widespread use of automatic translation, often without post-editing, contributes to the gradual erosion of the very notion of the “text” as a coherent and cohesive communicative unit. The paper also draws attention to pedagogical implications of these findings. Classroom experience in text analysis reveals that translation students often demonstrate perceptual difficulties similar to those of computer programs: they fail to relate word meanings and logical connectors to the broader context, do not perceive coherence beyond the paragraph level, overlook the extra-linguistic situation of the text, and experience difficulties in identifying intra-textual links. The study emphasizes the importance of deliberately teaching students to perceive texts as integrated wholes, using machine-translated texts as illustrative examples of the consequences that arise from violations of basic text-forming categories.
| Section | LANGUAGE AND CULTURE |
| DOI: | 10.47388/2072-3490/lunn2025-72-4-59-74 |
| Downloads | 198 |
| Key words | text; context; text-forming categories; coherence; cohesion; intra-textual links |