Gruzdev Dmitry Y.

The researchers note that translating technical texts into one’s native language is no easier a task than doing it in the opposite direction, particularly when the translator lacks expertise in the target area. The main difficulties that supposedly arise are translation of terms and clichés, as well as fitting the thought into an appropriate form in the target language. The authors suggest that ad-hoc corpora of texts in the native language be used to address them. The basis for this study became the experience, by all means successful, of utilizing the same type of resources, this time consisting of texts in the foreign language, in translations from the mother tongue. The researchers demonstrate strategies of employing corpora to overcome language difficulties and concurrently address the ever-pressing issue of optimizing the procedures for information retrieving from electronic translating tools. Particularly, they look into the benefits of corpus managers’ advanced search and wild card characters. Among other noteworthy results of this work are several tested and proven approaches to generating corpus queries based on the context analysis. They come in handy when everything else, i.e. dictionaries and other conventional translating tools, failed as sources of preliminary data for corpus searching.

The paper is focused on finding ways to exploit enhanced capabilities of corpus­managers, particularly the Regex function, utilizing wildcards in queries. The basis for this study derives from the experience, by all means successful, of utilizing this resource coupled with ad­hoc corpora by inexperienced translators to overcome translation difficulties. Since not entire potential was taken advantage of, the authors suggested that more complex functions could be useful for weathered translators, who would use them not to look for linguistic data but rather check their assumptions. To test the hypothesis the researchers carried out an experiment involving two groups of translators, experienced and beginners, who were asked to provide TAP (think­aloud protocol) [Someren, Barnard, Sandberg, 1994: 1] based translations of a passage. The analysis not only covering actual solutions and operations of the informants but also extrapolating their actions to potential outcomes is backed by examples. The authors came to the conclusion that Regex is a worthy thing given good language proficiency and subject savvy of the translator. Besides, the accuracy of queries, which comes inevitably as the specialist’s language proficiency improves and has an advert effect on the efficiency of linguistic data retrieving from a corpus, will be offset by wildcards. If used expertly, regular expressions provide flexibility making up for the lack of fuzzy string search [Bast, Celikik, 2013: 2] in corpus­managers.

The article dwells on an alternative approach to doing written translation powered by speech-to-text software. Triggered by a new trend in the professional community around the globe, it is mostly about replacing manual typing with a faster dictation coupled with special speech recognition technologies. The researches note that this entails swapping written translation with sight translation, a move fraught with greater difficulties due to time limitations and simultaneous perception, processing, translation and output of the target text. The paper focuses on major cognitive processes engaged in both types of translation. This has gone a long way in establishing hallmarks of the alternative approach and figuring out ways of optimizing the most complex aspects in sight translation to tailor it for a phase in written translation, providing faster input of the target text. The ultimate test of the compliance of the target text quality with the requirements for written translation comes in the form of an experiment encompassing adequacy, equivalence and pragmatic adaptation, provided by the conventional and alternative approaches. Based on the analysis of the two types of translation and experimental verification of the prompted hypothesis confirmed the feasibility of achieving better performance without compromising the quality of the target text by replacing written translation processes with some of those employed in sight interpretation. Though the approach is suitable for both young and experienced translators, the latter have better chances of achieving higher quality and productivity. Major constraints on the way of specialists with limited practice are the lack of sufficient number of ready templates and clichés, poor terminology base, and inadequate sight translation skills. Further attempts to expand the study of the speech-to-text powered approach to written translation may include dictation strategies and integration of the tool into the suite of modern translation software.