POST-EDITING MACHINE TRANSLATION OF USER MANUALS IN THE GERMAN–UKRAINIAN LANGUAGE PAIR: TYPICAL ERRORS AND THEIR CORRECTION
Abstract
The article focuses on analyzing typical errors in machine translation of user manuals from German into Ukrainian and identifies methods to effectively correct them when doing post-editing. The relevance of this subject is due to the rapid development of neural machine translation systems and the need to provide high-quality translations of user manuals, on which the proper use of household appliances depends. The aim of this case study is to identify errors in machine translation and to provide practical recommendations for correcting them. The research is based on German user manuals for Philips household appliances, as well as their machine translation into Ukrainian using the DeepL service. The study uses methods of analysis and synthesis, comparative, quantitative, and discursive analysis, as well as the method of modeling. By comparing existing classifications of machine translation errors, a generalized classification has been suggested that is relevant for identifying machine translation errors in user manuals related to omission, overtranslation, non-translation, and incorrect translation. This study reveals and classifies errors caused by incorrect translation. These errors include lexical, stylistic, orthographic, semantic, and grammar errors. It has been found that the most common causes of machine translation errors are terminology inconsistencies, literal translation, ignoring the stylistic norms of Ukrainian user manuals, the occurrence of lexical calques, and disregard for context. The paper emphasizes the importance of developing and using terminological glossaries, adhering to the norms of modern Ukrainian orthography, the stylistic features of technical documentation, and the applying of translation transformations during the post-editing of machine translation. The recommendations suggested can be used in the professional practice of translators of technical documentation and in training students in the skills of post-editing machine translation.
References
2. Calixto I., Liu Q. An error analysis for image-based multi-modal neural machine translation. Machine Translation. 2019. Т. 33. С. 155–177. URL: https://doi.org/10.1007/s10590-019-09226-9 (дата звернення: 29.07.2026).
3. Philips HU4801, HU4811 : інструкція з експлуатації / Koninklijke Philips N.V. 2019. 224 с. № 3000 039 83821.
4. Philips Sonicare : інструкція з експлуатації. Netherlands : Philips, 2022. 134 с. № 300005978843.
5. Philips. Handheld Steamer 5000 Series. Type STH5030/20 : інструкція з експлуатації. Drachten, Netherlands : Versuni Holding B.V., 2024. 80 с. Док. № 642001015050, ред. 03/06/2024.
6. Rodríguez-Rubio S. An analysis of Google Translate and DeepL translation of source text typographical errors in the economic and legal fields. Revista de Llengua i Dret, Journal of Language and Law. 2024. № 81. С. 88–105. URL: 10.58992/rld.i81.2024.4188 (дата звернення: 29.07.2026).
7. Salinas M.-J. V., Burbat R. Google Translate and DeepL: Breaking taboos in translator training. Observational study and analysis. Ibérica. 2023. № 45. С. 243–266. URL: 10.17398/2340-2784.45.243 (дата звернення: 29.07.2026).
8. Wiesmann E. Machine translation in the field of law: A study of the translation of Italian legal texts into German. Comparative Legilinguistics. 2019. Т. 37. С. 117–153. URL: 10.14746/cl.2019.37.4 (дата звернення: 29.07.2026).

This work is licensed under a Creative Commons Attribution 4.0 International License.
ISSN 
.png)



