Abstract:

Undoubtedly, machine translation has become an integral part of the translation process and translators prefer machine translation systems to a pen-and-paper approach, which is deemed rather obsolete. Although the quality of machine translation is high due to various innovations and developments which are constantly improving, a fully automated, top-quality machine translation does not yet exist. Machine translation is still not flawless; it can be helpful only along with post-editing. Implementation of post-editing into practice demands understanding new technologies and post-editor's training. Post-editor's competences differ from translator's competences; a good translator is not always a good post-editor since different activities require different skills and knowledge. If a new millennium offers new technological opportunities, it also needs to offer new educational opportunities. This fact should be taken into account when designing study programs and curricula in translatology since post-editing requires adequate education and training.

In our paper, we study the issue of post-editing in the context of the quality of machine translation in the direction English – Slovak in the publicistic texts. We introduce the current state of machine translation quality in the item of predicativeness and point out the most significant errors of machine translation in the given area. In our research, we use the adjusted error typology for the Slovak language devised by the Slovak linguist Juraj Vaňko (2017). His framework allows the evaluation of morpho-syntactic and syntactic-semantic relations - the areas with significant errors – into more details. The results show that errors in the category of ´agreement in gender´ correlate with errors in the category of ´tense´; the category of ´agreement in number´ correlates with the category of ´person´. The errors are related to the typological characteristics of the compared languages (English an analytical language, Slovak a synthetic language) and the nature of the examined texts (journalistic style). For a post-editor, stating the error rate and subsequent highlighting of the errors is crucial. When they recognize the issues and errors in machine translation, it is easier for them to detect and correct them.

Keywords:

Machine translation, post-editing, post-editor, skill, errors.