Abstract
<jats:p>This study undertakes a comprehensive analysis of audiodescription (audio commentary for visually impaired audiences) as an intersemiotic translation form through the lens of historical linguistics and cognitive-pragmatic theories. The aim is to identify and describe linguistic patterns underlying high-quality audiodescriptions, focusing on principles governing verbal description construction within its dependence on audio-visual content. We introduce the concept of “audiovisual proposition,” where sound-image sequences and accompanying textual components are viewed as interconnected elements — psychological subject and psychological predicate respectively. This framework enables us to interpret audiodescriptions not merely as passive replication but rather as active processes constructing meaning that facilitate holistic film perception by blind or partially sighted viewers. An official tiflocommentary for the Russian film “Moskva slezam ne verit” (1979) has been analyzed in this context. Typical errors identified include text fragmentation, use of single-clause sentences without accessible contextual cues, inappropriate past tense usage, and attempts at describing character traits instead of observable behaviors. Based on principles of psychological syntax, specific recommendations for correcting these segments have been proposed.</jats:p>