REVIEWING THE RESULT OF MACHINE TRANSLATION: A CASE FOR INDONESIAN TRANSLATION VERSION BY GOOGLE TRANSLATE AND IMTRANSLATOR

The research is intended to describe the translation version of a literary story that produced by the machine translation. This study used qualitative research. The research was conducted from a literary story of the storybook of Robin Hood: Level 2, retold by Liz Austin. The book was published by Pearson Education Limited in 2000 (the first edition). Research data and analysis were taken from the documentation of the translation results of different machine translation, that is Google Translate and IMTranslator. The result proves that both Google Translate and IMTranslator can help translators to do their works quickly. There are also some weaknesses from those MT as follows: (1) some versions are inaccurate and inappropriate according to the grammatical rules; (2) those MT can only translate on words for word and literal level, so the versions are looked inaccurate and inappropriate for some contexts; (3) those MT also cannot distinguish connotative or associative meanings; and (4) the results of MT serve only as pre-translation, therefore, they still need further discussion and adjustment by the human translators to achieve more acceptable, readable, and understandable results.


INTRODUCTION
Language is the key to introducing and understanding other people and cultures. Therefore, people need the media or even a way to cope with such barriers. One of the possible ways to solve those problems is translation. As Mutaqin and Sulistyawati (2020) informed that to communicate with one another or various countries with different language either directly or not, we need a translation. As time goes by, the world changes every time. The development of technology makes things easier, including in the field of translation. Information technology, which began to become popular in the late 70s, emerged to answer the challenges of the past through the term computer technology or electronic data processing or EDP (Electronic Data Processing). Information technology refers to the use of electronic equipment, especially computers to store, analyze, and distribute any information, including words, numbers, and images.
The development of information technology also can process, display, and disseminate information on learning in audio, visual, audiovisual, and even multimedia. This realizes what is called virtual learning. This concept develops surprisingly. So, it can package the conditions and reality of learning to be more attractive and provide adaptive conditioning to the learner where they are. The development of information and technology is also evolving in language learning. Various languages can be learned through information technology with the help of translation, or machine translation, precisely.

Translation
Most people assume that translation only a result, not a process. As Baihaqi (2017) stated that the translated texts which are read by TL readers are exactly the result of a translation practice; target readers are generally unaware of the complex process undertaken by translators when replacing the source language. Pym (2007) also stated that translation is the delivery of the source language into the target language as close as possible, and if possible, both have similarities which refer to the content and style of the source language. Those theories conclude that translation is the process of replacing the content and style of the source language as close as possible to the target language.
In transferring the source language, cultural aspects are closely related to the translation since it is not only deal with the message but also culture (Nababan, 2008). A translator is also a mediator of two different cultures, also known as asymmetric culture (Baihaqi, 2017); he must be able to process the result equivalently by using any available methods and procedures.
The translation methods and procedures related to how the translation process and result will be carried out. Baihaqi (2018) stated that translation procedures (also methods) as practical steps in analyzing and solving the problems in translation practices. Furthermore, Molina and Albir (2002) detailed the translation process into several translation procedures/techniques as follows: (1) borrowing, as a procedure that borrows words from the source language; (2) calque, it is a procedure by adjusting the target language structure; (3) description, as a procedure that replaces an expression by explaining in the form and function; (4) established equivalent, as a procedure that uses familiar expressions; (5) generalization, is a procedure that gives a common name; (6) particularization, is a procedure using a more concrete equivalent; (7) reduction, it is translating by reducing the meaning; (8) addition, it gives further clarity to the text; (9) deletion, it removes unnecessary information; (10) couplet, it uses of two translation procedures; (11) adaptation, it translates by matching the culture; (12) literal, it means translation using translating words; and (13) shift, it is a change in translation.

The Machine Translation
Machine translation, known as MT, is software that is currently customized to help the translation process. This is the attempt to automate all, or part of the process of translating human language (Arnold, Balkan, Meijer, Humphreys, and Sadler, 1996). Hutchins (1995) also stated that MT referred to a computerized system which produces translation from one natural language to another with or without the assistance of human translators. Regarding the development of MT, Wilks (2009)  Furthermore, Somers (1999) claimed that in the last ten years there has been a significant amount of research in MT within a 'new' paradigm of empirical approaches, often labeled as 'Example-based' approaches. The manifestation of this approach causes some surprise and hostility among observers to use different ways of working, but the techniques were quickly adopted and adapted by many researchers, often creating hybrid-systems.
Al-Deek, Al-Sukhni, Al-Kabi, and Haidar (2011) claimed that an automatic evaluation of two Well-known FOMT (Free Online Machine Translation) systems is Google Translate and IMTranslator. Google translate itself is a service provided by Google Inc., to translate parts of text or web pages in one language to another. For some languages, users are asked to provide alternative translations, such as for technical terms, which will be included for updates in the next process. Google uses it's as translation software. Based on the above discussion, therefore, the purpose of this study is to describe the translation results of a literary story that produced by the MT (without editing). The results of this study are expected to provide an overview that whether the MT can replace the human translator entirely, or it is only a tool and service to help the translators work.

METHOD
The research was conducted from the literary story of the storybook of Robin Hood: Level 2, retold by Liz Austin. This is one of the popular stories that attract the writers to have further discussion of it.

Results
The translation results from both Google Translate and IMTranslator can be observed as follows: But not everybody knows that he came from a rich family.
Tetapi tidak semua orang tahu bahwa ia berasal dari keluarga kaya. 5 And not many people know that Robin Hood was half-Saxon and half-Norman.

Discussion
Based on the result, the writers find some problems of accuracy, readability, and acceptability especially for word-for-word, grammatical, and contextual meaning. The translation for wordfor-word and literal occurred in most translation versions on Google Translate and IMTranslator.
Data 2 shows word-for-word translation. version is unacceptable to Indonesian grammatical system since it produces an incorrect result.
Next, data 12 shows that the Google Translate version is acceptable to the reader rather than IMTranslator; "Saya tidak punya anak laki-laki," kata Gamwell kepada Joanna., for Google Translate version, and "Aku punya anak-anak tidak," Gamwell dikatakan Joanna., for IMTranslator version. The version of IMTranslator is not only inappropriate to the structure of Indonesian Language but also it seems confusing to TL readers. Data 13,17,23,29,34,36,41,45,50,52  The last, it seems that IMTranslator tries to foreignize its translation version than Google Translate has. This can be observed at data 6, 11, 15, 21, 28, 31, 37, 39, 42, and 47. The term Sir in Sir George remains to be used in TL for IMTranslator versions; while Google Translate prefers to use Tuan as its TL version.
To sum up, based on 57 data above can be found that Google Translate output are more acceptable and understandable than IMTranslator has. This conforms to Al-Deek, Al-Sukhni, Al-Kabi, and Haidar (2011) study that based on ATEC reveal that Google Translate output were closer to human reference translations, relative to its counterpart IMTranslator. A similar result is also proved by Rafika, Miftahuddin, and Qutni (2019) that the result of IMTranslator tends to have more errors on grammatical aspects than the Google Translate has; and Ismail and Hartono (2016) that the most common errors of machine translation they were grammatical errors.

CONCLUSION
It can be concluded that both Google Translate and IMTranslator can help translators to do their works quickly. On the other side, the results also show some weaknesses resulted by those MT as follows: (1) some versions are inaccurate and inappropriate according to the grammatical rules; (2) those MT can only translate on words for word and literal level, so the versions are looked inaccurate and inappropriate for some contexts; (3) those MT also cannot distinguish connotative or associative meanings; and (4) the results of MT serve only as pre-translation, therefore, they still need further discussion and adjustment by the human translators to achieve