The Translation Quality of Facebook Translation
| Autor: | Arif Nurul Hakim, Andi Dian Rahmawan |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
business.industry
Computer science media_common.quotation_subject Literal translation computer.software_genre Translation (geometry) language.human_language Readability Test (assessment) Indonesian language Feature (machine learning) Quality (business) Artificial intelligence Calque business computer Natural language processing media_common |
| Zdroj: | JEdu: Journal of English Education. 1:203-208 |
| ISSN: | 2798-6543 2798-5164 |
| DOI: | 10.30998/jedu.v1i3.5981 |
| Popis: | This study aims to assess the quality level of the Facebook translation From English to Indonesian, what techniques are used and to determine the causes of the low quality of the automatic translation tool. It is hoped that this research can become a reference for netizens in understanding and concluding information that has been translated using Facebook's automatic translation feature to further be considered by developers of automatic translation technology to improve their products. This study compares the results of the Facebook translation with the results of the respondents presented with a qualitative descriptive method. The researcher conducted a test on 40 Facebook translation data involving 8 students of the English language education study program to assess whether the Facebook translation results were acceptable or still require re-translation from them. Furthermore, the researchers assess the translation quality and identified the translation techniques used by Facebook and respondents, and then those technique and quality were classified as propose by experts. From 48 techniques used by Facebook, found 48 data (100%) categorized as not accurate, not acceptable and low readability translations consisting of 37 data using literal translation (77,1 %), borrowing namely 9 data (18,7 %), calque namely 1 data (2,1%), and reduction namely 1 data (2,1%). Meanwhile, the translation techniques used by the respondents were more diverse. From 54 techniques used by Respondents, found 54 data (100%) categorized as accurate, acceptable and high readability translations consisting of 21 data using literal translation (38,8 %), adaptation namely 14 data (25,9 %), transposition namely 9 data (16,6%), borrowing namely 4 data (7,4%), description namely 2 data (3,7%), amplification namely 1 data (1,9%), calque namely 1 data (1,9%), modulation namely 1 data (1,9%) and reduction namely 1 data (1,9%). Facebook's tendency to often use Literal translation is believed to be the cause of the low level of quality because this technique is used as an early stage of translation and is called the lowest technique. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|