A Fuzzy Evaluation Model For SciTech Translation

In this thesis, existing translation assessment models are introduced firstly. Much attention has been paid to the quantitative evaluation models. During the process of introduction, the advantages and limitations of these methods are briefly discussed. Based on the existing quantitative evaluation models, a fuzzy evaluation model for SciTech (science and technology) translation is put forward.The quality of translations and the translation quality assessment are briefly introduced in Chapter 1. In Chapter 2, the existing qualitative and quantitative assessment methods are introduced in detail. There are some detailed discussions on their advantages and insufficiencies. Compared with literary translation, SciTech translation is more suitable for quantitative evaluation. In the third chapter, the characteristics of SciTech texts and their translations are analyzed in detail. They are characterized by objectivity, exactness and plainness, etc. In Chapter 4, the fuzzy evaluation method for SciTech translation is introduced. According to the fuzzy evaluation processes, the first step is to set up a parameter set. The parameters introduced in this model are the most relevant to SciTech translation because these parameters are decided after a detailed analysis of the characteristics of SciTech texts. In this thesis, the SciTech translation is evaluated from three aspects: microtext, macrotext and extratext.At the microtextual level, the evaluation concerns the quality of the translation of words and sentences. As to the evaluation of word, faithful transmission of the original word meaning is the first requirement. The second one is that words should be used correctly in the target text. There are mainly three aspects in the evaluation of the use of words: word collocation, the transference of word type and usage of function words. As to the evaluation of sentence, there are four aspects that must be examined. Firstly, there should be no grammar mistakes in the sentence. Secondly, the translation should faithfully and correctly transmit the sentence meaning. Thirdly, the sentence should be suitable for the SciTech text. Fourthly, the translation strategies adopted in the translation should be used appropriately. The translation strategies include addition, deletion, change of word order, transference of sentence structure, etc.At the macrotextual level, the overall quality of the translation must be examined. Three aspects must be taken into consideration in the evaluation of the overall quality of SciTech translation. The first aspect is logic. This is an important factor in theSciTech text which is highly logical. There are mainly three logic relations in SciTech text: concept, judgment and inference. The second aspect in the evaluation of macrotextual level is rhetoric. SciTech text is characterized by plainness, concise, objectivity, etc. The third aspect is cohesion. The cohesive relation must be transferred faithfully from the original. The cohesive devices must be used correctly in the target language.At the extratextual level, two parameters must be evaluated: the target readers and the purpose or function of the translation. Every writer has a target reader of his writing in his mind. It is also the case for the translator. The translation must be suitable for the target readers. Beside the parameter "target reader", the purpose of a translation is also a factor that must be examined.After setting up a parameter set, we should set up a grade set. In our model, there are four grades: excellent, good, acceptable and unacceptable. They are represented with A, B, C and D in the thesis. Each grade has its corresponding fuzzy set of the membership.The third step is the assignment of weight. The weight of each parameter is assigned according to the importance of it in SciTech translation.When evaluating each parameter, the evaluator should decide the grade of the quality of the translation according to fuzzy marking. So it is unnecessary for the evaluator to give an accurate score to each parameter. T...