Translating Phrasal Verbs from English to Arabic: Human vs. Machine Translation

Abstract

This study aims to compare between students’ and machines’ proficiency in translating phrasal verbs. So, a test of 20 phrasal verbs is given to 4th year students of English and Translation Department in the UST as well as to 5 machine Translation engines to translate them into Arabic. The students’ and machines’ translations are corrected from the viewpoint of being acceptable or unacceptable and then the students’ and machines’ performance are evaluated. Furthermore, this study focuses on the methods applied by the students and machines in transferring the given 20 phrasal verbs into Arabic. It identifies three translation methods frequently resorted to by the students and machines. These methods are literal, semantic and functional . In addition, the difficulties which are encountered by the students in the translation process of these 20 phrasal verbs are identified depending on a questionnaire of eight reasons of difficulties given to them. Based on the correction and analysis of the students’ translations, thirteen out of twenty students with a percentage of 65% pass the test, and the overall performance of the students is fair. Moreover, the most frequently used method is functional method with a percentage of 44.07 %. Regarding the students’ responses to the questionnaire, the difficulty that gets the highest "Agree" with a percentage of 45% is "The context is not enough to guess the meaning of the SL". Depending on the correction and analysis of the machines’ translations, it is found all the machines pass the test, and the overall performance of the machines is fair. The most frequently used method with a percentage of 63% is literal strategy. Comparing results of humans and machines, we find that the human performance is much better than that of the machines’ regarding the use of the functional method.