Analysis of Errors in Machine Translation from Roger T. Bell's Translation Process Model

Jianwei Zhu,Min Zhang
DOI: https://doi.org/10.6918/ijosser.202102_4(2).0062
2021-01-01
Abstract:There are enough literature reviews about machine translation, but the numbers of texts studied are not large enough, and there are very limited varieties of machine translation tools they have used. Taking the Chinese Government Work Report in 2020 as an example, this paper compares Google, Youdao, Baidu, Sogou and official translations from the perspective of syntax and pragmatic structure by using Roger T. Bell's translation process model. It is found that new words are not introduced into the Frequent Lexis Store in time, and the organization of sentence structure is not appropriate, and pragmatic awareness is weak through the process of machine translation. In the same text, Baidu translation has the most words, and Sogou translation has fewer grammatical problems, and Youdao translation has better pragmatic performance, and Google translation has no outstanding highlights.
What problem does this paper attempt to address?