a comparative study of the english versions of persian homonyms as translated by google translate and microsoft translator

a comparative study of the english versions of persian homonyms as translated by google translate and microsoft translator

language as the biggest and most important tool for making communication is of different varieties. no one can claim they know and understand all the languages in the world. so, rendering concepts from one language into another plays an important role in making an easy and speedy communication possible. translation machines have helped this event very much since their first appearance in modern technological world.

this study investigates translation quality of two translation machines, google translate and microsoft translator, as the commonest translation machines to find differences between their operations in the area of literature and show which one performs better. to meet the objectives of the study, the researchers concentrated on hafez’s divan and his ghazals. they reviewed and studied 201 ghazals of hafez and searched for homonyms in them.

thus, 481 homonyms were found. in the next phase, google translate and microsoft translator were fed by persian verses including the homonyms to get their english renderings. the machines’ outputs were subjected to both human evaluation and the bleu software scoring. the results were analyzed through paired samples t-tests. the findings showed that though a significant difference existed between human and machine translation assessment, no statistically significant difference existed between the translation quality of the machines as evaluated by human or bleu.