An Empirical Study Of The Reliability And Validity Of Testing Translation Quality By Back-translation

Fully approving of the effectiveness of testing translation quality by back-translation, scholars on translation studies have widely adopted this approach to researches in this field. However, they have not attached due attention to the reliability and validity of this method, especially the possible influence of the back-translator’s subjectivity on the back-translating process. Therefore, this paper intends to describe how the back-translator’s subjectivity influences the reliability of translation quality test from the perspective of explicitation theory. It also attempts to explore possible ways of increasing the validity of testing translation quality by back-translation under the guidance of relevance theory.The first experiment aims to explain the relationship between the back-translator’s subjectivity and the reliability of translation quality test by comparing the back-translations made under two conditions:Condition A is that back-translators do not know it is back-translation in essence; Condition B is that back-translators do know it is back-translation in advance. Thus, the translation material had better not contain obvious hints reminding the participants of a Chinese-English translated version. The excerpt of the English version of Fortress Besieged (translated by Jeanne Kelly and Nathan K. Mao) is an ideal experimental material. Though it has only300words, this excerpt is well organized in plot and idiomatic in expression. Most importantly, it shows few traces of translation void of Chinese names (Fang Hung-chien), unique Chinese objects (rickshaw), or proper nouns (War of Resistance). The result shows that the enlargement of the back-translated text is smaller and the explicitation of personal pronoun subjects is more manifest than those made under Condition A. It proves that the back-translators can better play their subjectivity under Condition B, for they can understand the translation better, use less explanatory vocabularies and reduce redundancy in their renderings. Therefore, their back-translations enjoy higher reliability of testing translation quality.In line with the above findings, the second experiment carries out comparative analyses of the back-translations made under Condition B. The material is an excerpt of two complete English versions of Hong Lou Meng (translated by David Hawkes, Yang Xianyi and his wife Gladys Yang). The participants are divided into two groups back-translating Hawkes’and the Yangs’ versions respectively. This classic is regarded as "an encyclopedia of Chinese culture", and the two complete English versions have aroused incessant interest of scholars on translation studies. Moreover, the chosen excerpt has been discussed by different scholars from various angles. Therefore, this experiment hopes to evaluate the translation quality of the two versions from the new perspective of back-translation. It can be concluded that both Hawkes’and the Yangs’translations have their own advantages of keeping inferential clues. Specifically speaking, the two versions have born a high resemblance to the original text on two special Chinese sentence patterns:sentence with serial verbs and simple sentence with no conjunction. Thus, their back-translators can effectively infer the original author’s intention and seek representation of the original text. With respect to semantic clues from three aspects, namely, unique Chinese objects, Chinese names and Chinese idioms, Hawkes’translation has kept them deliberately either to create contextual assumptions or to adjust aesthetic expectations for his target readers. But his consideration fails to take due effect but makes its back-translators confused and bewildered. The Western readers have also expressed similar feelings in the survey of evaluating the two versions. As for the Yangs’translation, though the semantic clues are not manifest, both the back-translators and the Western readers can grasp the gist of the original author’s intention well.Theoretically, attempts have been made in this study to clarify that the back-translating process is an ostensive-inferential communication just the same as translation. The back-translator is playing dual roles during the process:a TL reader trying to infer the original author’s communicative intention from the translator’s inferential clues; and a translator needing to make the inferred information ostensive in the SL.Practically, the findings from this study justify the practice of back-translation in translation quality test. Moreover, the analytic approach to evaluating inferential clues sheds new light on the comparative study between Chinese and English. Finally, relevant findings may also provide some insights for translation teaching.