Machine Translation: Linguistic challenges that arise in the translation of journalistic texts
MetadataShow full metadata
A quality translation involves a process that goes beyond matching words, phrases, and sentences between two languages. Machine Translation (MT) systems use applications to translate text from one language to another without the intervention of a human translator (TH), resulting in poor translations. The purpose of this research is to identify the most frequent errors in translations produced by MT systems from English to Spanish. Through a comparative analysis between MT and HT, linguistic limitations will be analyzed in terms of grammatical, semantic, and lexical issues. The texts for this evaluation will be obtained from online news sources, which will be translated through the automatic translation services of Google Translate (GT) and Bing Translator (BT), leading companies in MT. Both services use Neural Machine Translation (NMT) systems, which mimic the learning process of the human brain. These translation services claim that computer can learn and "think" on its own without human intervention. NMT systems promise simpler techniques that can process translated sentences data to “train” the translation model between two languages. These systems appear to be very promising for the current demand of the translation industry. However, translations produced by MT do not have the quality control required in most professional translation practices.