Contrasting and Analyzing Machine and Human Translation

The quality of machine translation is an important part in machine language competence. Currently, there are still many differences between machine translation and human translation. The quality of machine translation still could not reach the quality of human translation. We take a Chinese literary text, Mo Yan's Red Sorghum, as the material to translate, and used LIWC2015 to analyze the differences between machine and human translation. Also, we make a linguistical attribution analysis to find out the reasons behind the errors machines make. For the engines of machine translation, we choose Baidu, Youdao and Google as machine translate engines. By analyzing the results from LIWC2015 and error analysis, we find that compared to human translators, machine translate engines tend to use simpler words and often fail to recognize discoursal connection; the errors machine translate engines make are usually caused by long sentences, Chinese names, encyclopedic knowledge, unique expressions and constructions in Chinese and literary expression.