UDC

81ʼ322.4:[811.511.11+811.161.1]

DOI

10.37482/2687-1505-V421

Authors

Anna V. Merzlaya
Postgraduate Student, Translation Studies and Applied Linguistics Department, Northern (Arctic) Federal University named after M.V. Lomonosov (address: prosp. Lomonosova 4, Arkhangelsk, 163002, Russia).
e-mail: aannaa@yandex.ru, ORCID: https://orcid.org/0009-0008-7109-3351

Abstract

This article presents a systematic description of translation failures resulting from neural machine translation of Finnish-language meteorological media texts (forecasting a cold front and negative temperatures) into Russian. The importance of the topic is determined by the increasing attention of meteorologists to phenomena related to climate change and weather anomalies as well as by the interest of IT specialists in optimizing neural machine translation systems that can significantly accelerate the process of interlingual communication. The key aim of the study is to identify and systematize translation failures that occur when translating Finnish-language meteorological texts about cold weather published in the online version of the newspaper Iltalehti as well as attributes of the cold used in these texts. First, micro-contexts were selected by means of continuous sampling, then the selected meteorological texts were translated using the online service DeepL Translate, and finally, by means of component and contextual analyses, the main types of translation failures were identified and the reasons for their occurrence were described. The study demonstrates that the greatest number of failures when translating verbalizations of the cold include omitting cold sensations, i.e. the seme of the cold that is present in the semantic structure of the Finnish lexeme is not rendered in the Russian translation. Moreover, translation failures can arise from the choice of a lexical unit inadequate to the context or use of proper names and nonce words whose rendering requires finding original translation solutions and typically cannot be realized through neural machine translation. The results of the study can be used to optimize neural machine translation systems as well as to teach translation and post-editing of translations of meteorological media texts.

Keywords

verbalization of cold perception, meteorological texts, Finnish language, neural machine translation, translation failure, interlingual communication

References

1. Miftakhova R.G., Morozkina E.A. Mashinnyy perevod. Neyroperevod [Machine Translation. Neural Translation]. Vestnik Bashkirskogo universiteta, 2019, vol. 24, no. 2, pp. 497–502.
2. Egorova A.Yu., Zatsman I.M., Nuriev V.A. Ekspertnaya otsenka mashinnogo perevoda: klassifikatsiya oshibok [Expert Evaluation of Machine Translation: Error Classification]. Sistemy i sredstva informatiki, 2021, vol. 31, no. 3, pp. 144–157.
3. Goncharov A.A., Buntman N.V., Nuriev V.A. Oshibki v mashinnom perevode: problemy klassifikatsii [Machine Translation Errors: Problems of Classification]. Sistemy i sredstva informatiki, 2019, vol. 29, no. 3, pp. 92–103.
4. Popović M. Error Classification and Analysis for Machine Translation Quality Assessment. Moorkens J., Castilho S., Gaspari F., Doherty S. (eds.). Translation Quality Assessment. Machine Translation: Technologies and Applications. Vol. 1. Cham, 2018, pp. 19–158. https://doi.org/10.1007/978-3-319-91241-7_7
5. Daems J., Vandepitte S., Hartsuiker R.J., Macken L. Identifying the Machine Translation Error Types with the Greatest Impact on Post-Editing. Front Psychol., vol. 8. Art. no. 1282. https://doi.org/10.3389/fpsyg.2017.01282
6. Barieva Yu.Yu., Abramicheva E.N. Difficulties in Rendering Perception Nominations. Ivantsova Yu.A., Rudenko N.S. (eds.). LinguaNet. Sevastopol, 2019, pp. 81–85.
7. Yaroshenko P.V. Modeli perevoda sensornoy leksiki (na materiale korpusa mnozhestvennykh perevodov) [The Translation Models for Sensory Lexicon (a Case Study of the Multiple-Translation Corpus)]. Available at: https://conference-spbu.ru/conference/50/reports/17918 (accessed: 14 January 2025).
8. Alekseeva L.M. Perevodcheskaya oshibka vs perevodcheskaya neudacha [Translation Error vs Translation Failure]. Lingvisticheskie chteniya – 2006. Tsikl 2 [Linguistic Readings – 2016. Series 2]. Perm, 2006, pp. 3–9.
9. Kharchenko V.K. Lingvosensorika: Fundamental’nye i prikladnye aspekty [Sensory Linguistics: Fundamental and Applied Aspects]. Moscow, 2012. 216 p.
10. Banartseva A.V. Yazykovye perevody: chelovek vs mashinnyy perevod [Language Translations: Man vs. Machine Translation]. Vestnik nauki i obrazovaniya, 2008, vol. 1, no. 8, pp. 50–53. Available at: https://cyberleninka.ru/article/n/yazykovye-perevody-chelovek-vs-mashinnyy-perevod/viewer (accessed: 19 November 2024).
11. Llitjós A.F., Carbonell J.G., Lavie A. A Framework for Interactive and Automatic Refinement of Transfer-Based Machine Translation. Proceedings of the 10th Annual Conference of the European Association for Machine Translation (EAMT). Budapest, 2005, pp. 87–96.
12. Berezina Yu.V., Baykova O.V., Skurikhina O.V. Lingvisticheskiy analiz leksiko-grammaticheskikh oshibok mashinnogo perevoda [Linguistic Analysis of Lexico-Grammatical Errors in Machine Translation]. Nauka. Obrazovanie. Obshchestvo: tendentsii i perspektivy razvitiya [Science. Education. Society: Trends and Prospects for Development]. Cheboksary, 2020, pp. 196–197.