SUMMARY
Traditionally, translation is viewed as a reliable shield over linguistic diversity, one of the ways to ensure a target language survival. However, translation is also reported to distort a translated language due to introducing ‘the third code’ (Frawley, 1984) features. These “deforming tendencies” (Berman, 1985) destroy the translated language by erasing its natural pattern and by adding there a bundle of alien features that cause its lexical, syntactical, and stylistic deficiencies. The current study is aimed at detecting those destructive features treated in translation studies as “translation universals” (Chesterman, 2004). To this end, a psycholinguistic analysis was held to establish the use of language which is not the result of intentional, controlled processes and of which translators may not be aware. These subliminal translation-inherent processes can be traced in the use of function words that encode procedural meaning. Relevance Theory (Wilson & Sperber, 1993) explains a conceptual-procedural distinction as a major distinction made between two types of linguistically encoded information. Conceptual information expressed by content words is viewed as encoding concepts whereas words with procedural meaning contribute to the derivation of implicatures, certain ways of processing propositions. Discourse connectives, conjunctions, prepositions, particles, pronouns, modal words constitute that group of function words with procedural meaning. To uncover certain variations in the use of these linguistic units, a parallel English-Ukrainian corpus made up of an 8,000-character excerpt from Franny by J.D. Salinger, its professional translation, and forty novice translators’ target versions, was compiled.The corpus data were processed by Textanz and SPSS computerized tools.The results of the psycholinguistic analysis proved that the Ukrainian versions as contrasted to the original text contained the following S-universals: implicitation expressed through the shortage of discourse markers of global coherence, simplification due to the lack of personal pronouns, decreased mean number of words per sentence, and greater number of sentences; normalization embodied in vernacular network impoverishment due to the decreased amount of pragmatic markers and fillers, explicitation due to higher lexical variety and density rates, and rationalization as a result of abundant marking of discourse relations.Conclusions. Taken together, these findings have significant implications for the understanding of how procedural information processing by novice translators is manifested in translation.