Effect Of The Representation Of Prefabricated Chunks On The Lexical Selection In Chinese To English Translating Process

Recently, the prefabricated chunks have become a hot research topic abroad and in China. Previous studies were mainly on the relationships between the use of them and second language competence, such as listening, speaking, reading and writing. However, little attention has been given to the mental representation of prefabricated chunks and in particular its relationship with the lexical selection in translation. Therefore attempts have been made to introduce the representation of prefabricated chunks into the study of the translating process to find whether there is any correlation between the representation of prefabricated chunks and the quality of lexical selection in C-E translating process.The representation of prefabricated chunks in the present study is defined as"multiword combinations, continuous or discontinuous, which is stored and retrieved as a whole from memory at the time we use them."They are classified into the four types according to Nattinger & DeCarrico (2000), Chunk1 Polywords, Chunk2 Institutionalized expressions, Chunk3 Phrasal Constraints, and Chunk4 Sentence Builders.To find out the correlation between the representation proficiency of prefabricated chunks and the quality of lexical selection, two experiments have been run within the theoretical framework of connectionist representation of prefabricated chunks in information-processing model of translation, which was specially developed for the current research. A lexical decision task (LDT) was used in Experiment 1 to assess the proficiency of representation of four types of prefabricated chunks in Chinese EFL learners by the two indices: mean response latency (with millisecond accuracy) and accuracy rate. The word translation task was conducted in Experiment 2 to test Chinese EFL learners'quality of lexical selection in C-E translating process in terms of fluency (by mean response latency in translation) and accuracy (by translation score).Data got from the empirical study are used to answer the following four questions:Q1. What are the main features of mental representation of prefabricated chunks in Chinese EFL learners?Q2. What are the possible features of lexical selection quality characterized by Chinese EFL learners in C-E translating process?Q3. What is the relationship between the Chinese EFL learners'proficiency of representation of prefabricated chunks and their quality of lexical selection in C-E translating process?Q4. To what extent can representation of different types of prefabricated chunks predict the Chinese EFL learners'quality of lexical selection in C-E translating process? With the statistical analysis of SPSS 13.0, the following major findings are drawn:Firstly, none of the Chinese EFL learners are good achievers in the representation proficiency of the prefabricated chunks. There are differences within high and low-level groups in the developmental patterns of representation of prefabricated chunks. As for the high group, the developmental route is: Chunk1> Chunk3> Chunk4> Chunk2. For the low group, the route can be summarized as Chunk1> Chunk4> Chunk3 > Chunk2.Secondly, as a whole, Chinese EFL learners'lexical selection quality in the translating process is far from satisfactory. The cross-group comparison indicates that there exist significant differences between the two groups in terms of fluency and accuracy of lexical selection in C-E translating process.Thirdly, Chinese EFL learners'proficiency of representation of chunks has a negative correlation with their mean response latency in translation. The value of correlation is high. The representation proficiency has a positive correlation with the translation score but the value is moderate.Fourthly, in the high-level group, the representation of Chunk2 has been proved to be a strong predictor of the fluency of lexical selection, being able to account for 57.4% variance of the fluency. The representation of Chunk3 has the strongest predicating power for the accuracy of lexical selection, accounting for 71.1% variance of accuracy. It is shown that, as for the low-level group, the representation of Chunk1 has the strongest predicting power for both the fluency of and accuracy of lexical selection, accounting for 42.3% variance of fluency and 58.9% variance of accuracy.Based on the theoretical framework the author analyzes the results and finds that the stable interconnections, connectionist representation of prefabricated chunks in the long-term memory can suppress the spreading activation and retrieval of non-target words, ensure that the parallel distributed process of the lexical selection takes place between the words set and within the word set at the same time. Thus, the fluency and accuracy of lexical selection in translating process are guaranteed.Theoretically, this study tries to describe a developmental route of the representation of four types of prefabricated chunks and its'impact on the quality of lexical selection from the low- level state to the high-level state (See Figure 6.2), and tentatively proposes a psychological model of lexical selection mechanism in translating process with the features of representation of prefabricated chunks (See Figure 6.3, 6.4). Pedagogically, findings got from this research have some implications for the teaching of vocabulary, chunks and C-E translation.