Performance of Google Translate, Microsoft Translator, and DeepL Translator: Error Analysis of Translation Result

Abstract

This research simulates the use of Google Translate and Microsoft Translator in translating Indonesian into English abstract and analyzes errors of translation results in the aspect of grammar and spelling using Grammarly Premium. This research is descriptive qualitative. The result analysis shows several writing issues (errors) of translation results using Microsoft Translator (MT), Google Translate (GT), and DeepL Translator (DT) in the aspect of clarity, correctness (correction), and delivery. First, there are 26 writing issues from MT including 1) 4 clarity (4 unclear sentences), 2) 21 correctness (2 comma misuse within clauses, 4 punctuation in compound/complex sentences, 5 wrong or missing prepositions, 3 incorrect verb forms, 2 determiner use, 2 conjunction use, 2 confusing words, 1 improper formatting), 3) 1 delivery (1 incomplete sentences). Second, there are 25 writing issues from GT including 1) 6 clarity (4 unclear sentences and 2 wordy sentences), 2) 17 correctness (1 misplaced word or phrase, 1 text inconsistencies, 1 comma misuse within clauses, 4 punctuation in compound/complex sentences, 3 wrong or missing prepositions, 3 incorrect verb forms, 2 conjunction use, 1 confusing word, and 1 determiner use.). 3) 2 delivery (2 incomplete sentences). Third, there are 10 writing issues from DT including 3 comma misuse within clauses, 1 incorrect noun number, 1 faulty subject-verb agreement, 1 wrong or missing preposition, 1 improper formatting, 1 misuse of modifiers, 1 conjunction use, and 1 confused word. Online translator including MT, GT, and DT may help in translating foreign language texts, but it does not always provide accurate translations. Thus, whether consciously or unconsciously, each translation machine contains its traits, whether directly or indirectly, human participation cannot be separated from machine translation technology. However, no computer translator can provide translations with the same precision as humans. It means that they are still a machine that is not able to match human capabilities equipped with feelings and thinking processes that describe them.