Machine Translation vs. Human Translation: Semantic Distinctions in English-Croatian Translations

Translation, a discipline that is a vital part of the knowledge and the study of a particular language, at the beginning of the last century was only associated with human knowledge and ingenuity. Nowadays, owning precisely to the same ingenuity, it is not just attributed to humans but also to the machines in a substantial way. In this paper, we will give a brief overview of translation as a discipline and its history before focusing in detail on machine translation, as one medium of translation. The human translation (HT) will be compared to a machine translation (MT) of the same texts taken from different nonliterary genres: administrative, criticism, blog, and academic, to provide some answers to the questions of whether or not machines have reached the level of proficiency that would make human translators and thus human translations obsolete in comparison, whether machine translation has been able to improve upon its original issues and in what way a human translation differs from a machine one. We will analyze these texts at a semantic level, i.e., our efforts will be focused on comparing how machine translation software deals with notions such as polysemy, context, and idioms, thus covering just a fraction of the vast area that is covered not only by machine translation but also by translation as a whole. The findings of this paper are in no way definitive given the small sample size of texts and also the fact that only a single machine translation software (Matecat) was used in the process