Reem Bazzal, and Nour Onies

Abstract In their text Evaluation of Machine Translation System and Human Translation with Literary texts (2013), Reem Bazzal, and Nour Onies intend to discuss the proper use of available online machine translation technologies in literary translation and provide an objective view of the suitability and reliability of these systems in translation. To illustrate more, practical text has been submitted and it was translated by both Machine Translation and Human Translation. They evaluated the work of such technologies and of human translation from four different aspects: semantics, syntax, morphology and comprehension and concluded that although Machine translation results are not that reliable, they can help in producing a well understood text. This paper is set out to verify this belief. Comprehension For a fair comparison of the performance of different online MT systems and Human Translation on literary texts, they needed appropriate text that the MT systems can evaluate. The text they choseThe Raven by Edgar Allan Poe is made up of 406 words in its English version. It is therefore a reliable test set for examining the translation quality of literary texts by MT systems and human translation and to produce substantial statistical analysis for their performance in quantitative evaluation. After translating the text by both human translation and machine translation which is google translate, we concluded the following: Semantics: When it comes to semantics, the meaning or the interpretation of a word or sentence we noticed that there are differences and similarities between the human translation and the machine translation. For example, in human translation, the expression midnight dreary was translated into ليلةٍ مقفرة while google translate translated it into ليلة كئيبة. Another example is the word pondered which in human translation meant تأملت while in machine translation it was translated to فكرت. “Rapping at my chamber door" was translated into “يقرع علىبابمخدعي" in human translation while with google translate it was translated to "موسيقىالرابفيبابحجرتي Syntax: Concerning the word order (syntax): the human translation is better than machine translation. The word combination in human translation is respected and well organized. It is more accurate and follows the grammatical rules: subject-verb and we can also notice the proper use of words such determiners, modifiers, and complements while in machine translation a majority of phrases started with the verb or the verb phrase instead of the noun. For example: Source sentence: And the silken sad uncertain rustling of each purple curtain Thrilled me Machine translation: وحريريسرقةمؤكدحزينةلكلستارارجواني Human translation: الحفيفُالحريريُّالحزينُوالغامضُلكلِّستارةٍأرجوانية أثارني Source sentence: While I nodded, nearly napping, suddenly there came a tapping Machine translation: بينماكنتضربةرأس،القيلولةتقريبا،فجأةهناكجاءالتنصت، Human translation: وعندماسقطرأسيعلىصدريوأوشكتُأنأغفو فجأةًجاءَالطرقُ Morphology: Concerning the morphology, machine translation has a problem when it comes to translation related to Inflectional morphemes, the exact number, gender of the subject and the exact verb. When translating from English to Arabic, the agreement between grammatically linked items was often missing which led to some sentences that can be hardly understood. While human translation has proved its efficiency in giving the exact verb tense number, possession, and comparison. For example: Source text: While I nodded, nearly napping, suddenly there came a tapping Human translation: وعندماسقطرأسيعلىصدريوأوشكتُأنأغفو Machine translation: بينماكنتضربةرأس،القيلولةتقريبا،فجأةهناكجاءالتنصت، Another example on the wrong morphological interpretation that was clearly stated in the translation process: Source text: As of someone gently rapping, rapping at my chamber door. Human translation: وكأنأحداًبرفقٍيقرعُ...يقرعُعلىبابِمخدعي Machine translation: اعتبارامنبعضواحدموسيقىالراببلطف،موسيقىالرابفيبابحجرتي After comparing the two methods of translation we noticed that the target texts in the machine translation and despite the mistakes can be understood. But the use of grammar and punctuation is not correct and the sentences are sometimes lengthy and complex. Also, MT produces some sentences that are unfortunately far from the meaning given in the source text. As for human translation, the texts we well understood, the meaning was clearly stated in the target language and the use of grammar and punctuation is correct. Take the following sentences as examples: Source sentence: Once upon a midnight dreary Machine Translation:مرةواحدةعندمنتصفالليلالكئيب Human Translation:فيليلةٍ مقفرةٍعندمنتصفالليل Source sentence: Over many a quaint and curious volume of forgotten lore Machine Translation:أكثرمنوحدةتخزينالعديدغريبةوغريبةمنالعلمالنسيان Human Translation مررتُبعديدٍ منالمعتقداتِالقديمةِالغريبة التيغطاهاألنسيان Source sentence: distinctly I remember it was in the bleak December Machine Translation:بوضوحتذكرت،أنهكانفيديسمبرالقابض Human Translation:أتذكّرُبوضوحٍ إنه ديسمبرالكثيب Source sentence: the rare and radiant maiden whom the angels name Lenore Machine Translation: النادرةالمتألقةالنادرةالذينلينوراسمالملائكة Human translation: لتلكالفتاةِ الرائعةِالمتألقة التيأسمتهاالملائكةُُلينور Recommendations It is important to include in MT engines special lexicon domains that can be used in literal translation, which offers quite a wide variety of lexicons if one accesses the MT sites directly. It is necessary to take into consideration that on-line MT engines are aimed at helping users deal with ephemeral literary texts and they help communication in many situations. 1. It is important to understand that machine translation has definitely improved the work of journalistic translation whether by the use of blogs, emails or portables like new Samsung phones (S. Translator) and IPhone phones( I translate) and the intelligent use of them led to the focus of the best human efforts where they should be. Machine translation like blogs and iPad and S. translator must be used an alternative to the expensive human translation that takes time and is usually unavailable when it is needed for communicating quickly and cheaply with people with whom we do not share a common language.