The rapid development of machine translation (MT) tools, particularly Google Translate (GT), transformed how people overcome language barriers in both everyday communication and academic contexts. Despite its popularity and accessibility, the accuracy and naturalness of GT translations remain questionable, especially in specialized fields such as academic writing. In 2021, three people – Menik Winiharti, Syihabuddin and Dadang Sudana – conducted a study to examine how effectively Google Translate renders Indonesian academic texts into English. The research focused on assessing not only how GT transfers meaning, but also how naturally it reconstructs the forms and structures of English sentences.

The study aimed to determine how Google Translate processes Indonesian lecturers’ academic writing and to what extent the resulting English translations are accurate and natural. It specifically analyzed academic papers written by lecturers from three different disciplines: Management, Japanese Literature, and Mathematics. Only the discussion sections of these papers were examined, as that part typically reflects the author’s original thought and expression. From each paper, two sentences were selected randomly, resulting in a total of six sentences for detailed analysis.

The researchers employed a descriptive–qualitative approach, focusing on both syntactic and semantic aspects. In the syntactic or form-based analysis, the structure of the source language (Indonesian) was compared to that of the target language (English) to see whether the translation produced natural English constructions. In the semantic or meaning-based analysis, the researchers evaluated how accurately the original message was conveyed in English by looking at both grammar and translation accuracy.

The findings revealed that Google Translate primarily applies a literal translation approach, closely following the structure of the Indonesian source text. While this method often preserves grammatical alignment between the two languages, it tends to produce sentences that are syntactically correct yet unnatural in English. Most translations maintained equivalent forms between Indonesian and English, but meaning inaccuracies were found in several areas. For example, certain lexical choices were imprecise (‘penduduk’ translated as ‘population’ instead of ‘people’), while idiomatic and contextual expressions were often mistranslated (‘makanan siap saji’ as ‘prepared foods’ instead of ‘fast food’). These issues highlight GT’s limited ability to interpret context and nuance.

The study emphasized that while Google Translate can deliver a general understanding of academic texts, it still lacks sensitivity to linguistic subtleties such as idiomatic usage, contextual meaning, and discourse-level cohesion. Some inaccuracies were also attributed to the source texts themselves, which were not always well-structured in Indonesian. The researchers noted that a poorly constructed source sentence could further complicate GT’s processing, leading to less accurate translations. Nevertheless, GT occasionally demonstrated effective syntactic adaptation, particularly when adjusting Indonesian noun phrases to fit English grammar conventions.

In conclusion, the research confirmed that Google Translate remains a valuable but imperfect tool for academic translation. It performs reasonably well in conveying the general message of Indonesian texts but still requires human intervention for refinement. The authors recommend that Indonesian lecturers and academic writers use GT as an aid rather than a substitute for human translation. They also stress the importance of reviewing and editing GT outputs to ensure linguistic accuracy and stylistic appropriateness. For language educators and institutions, the findings suggested integrating translation awareness and post-editing skills into academic writing training programs to enhance the quality of English publications by non-native speakers.

Winiharti, M., & Sudana, D. (2021). The English google translation of Indonesian lecturer’s academic writing: A preliminary study. Journal of Language and Linguistic Studies, 17(2), [706]-719. https://search.informit.org/doi/10.3316/informit.215875937433823

Article written by Rafli Izzudin B26.