A Corpus-based Study On The Features Of CFL Learners’ Use Of Chunks In Spoken Chinese

A number of large-scale corpus studies show that there exist many highly patterned, fixed or half-fixed prefabricated units, which are called chunks, in a language. Chunks, reflecting the conventional use of language, are stored and retrieved completely from memory, and they play a central role in achieving native-like fluency and selection. The development of native speakers’ ability to use chunks reflects the progress of memory-based holistic system of language process. Furthermore, chunks are important for native speakers to achieve the balance between creative and conventional expressions.The role of chunks is increasingly emphasized in second language acquisition (SLA). Some scholars investigated how second language (L2) learners of English used chunks and generalized some features from them based on corpus data. However, little empirical research has been conducted on how learners of Chinese, as a second/foreign language, employ chunks so far.The learner corpus in this study was built on self-collected oral outputs. It consists of two parts:the oral test sub-corpus and the improvised dialogue sub-corpus. The first sub-corpus contains the transcripts of the oral production of CFL learners at three different proficiency levels:elementary, intermediate and advanced. The second sub-corpus mainly includes the transcripts of the high level CFL learners, some of whom are studying a master’s degree in Chinese. The corpus of native speakers was built on the materials extracted from National Broadcast Media Language Corpus. Based on it, the chunks in different corpora were selected by using the automatic extraction tools and manual intervention. Adopting a comparison approach this study investigated the chunks used in CFL learners’ spoken Chinese vertically and horizontally through both quantitative and qualitative analyses.This study mainly investigates the following research questions:(1) what’s the development trend for the chunks used by learners of three levels? What are the specific characteristics in quantitative and qualitative terms? (2) What are the distinctive features of chunks used in the high level learners’ spoken Chinese, compared with native speakers’? What are the main factors influencing the features? (3) Are the conclusions generated from CFL learners consistent with those of EFL learners? Do the results support the dual-mode system of language processing? To solve the questions above, this study collected a large number of data, and applied the Log-likelihood test to determine the distinguished features of chunks used by the CFL learners. In addition, this study explored the qualitative features by focusing on the usage of specific chunks.This study has mainly the following findings:(1) The developing trend of chunks used by CFL learners didn’t show a linear correlation with their Chinese proficiency levels.1) The quantitative results showed the types and tokens of chunks used by different learners exhibited in a V-type and A-type respectively. The performance of elementary learners was significantly higher than that of the intermediate and advanced learners in the use of chunks consisting of 3-6 words. The proportions of chunks in different forms were getting closer to the level of native speakers, with the increasing level of L2 learners in Chinese.2) The qualitative results revealed that the use of interpersonal and discourse-organizing chunks developed substantially from elementary to intermediate levels. However, some chunks used by the latter were semantic and functional inappropriately which led to more over-used and misused chunks. The chunks used by advanced learners decreased in number but increased in accuracy. In the expressions on specialized topic, the advanced learners used more chunks referring to names, but the matched adjective and verb chunks didn’t increase accordingly. As a result, the expressions of advanced learners were not appropriate and idiomatic enough.(2) The chunks used by high level learners were distinctively different from those of native speakers.1) The tokens of two-word chunks used by high level learners were significantly more advanced than native speakers, but poorer on the types of two-word chunks as well as the types and tokens of 3-6 words chunks.2) The high level learners over-relied on some narrowly -focused chunks with high frequency. Among their overused and idiosyncratic chunks, there existed a number of chunks with similar forms. Furthermore, those high level learners seldom used the chunks reflecting the conventional use of native speakers, or those that have no direct semantic contributions.3) The features of chunks used in high level learners’spoken Chinese were influenced by the Chinese rhythmic and structural characteristics, the frequency effects in SLA, the sequence in language acquisition which is from concrete to abstract, and the high pressure of immediate coding in oral production jointly.(3)The development models on the dual-mode system of language process between L2 learners and native speakers were different. For native speakers, after going through the second stage which was dominant by analytically-processed language, achieved the third stage in which the holistically-processed language increased again. However, the trend of chunks used in learners at three levels and the features used by the high level learners indicated that the development of holistically-processed language for L2 learners usually stayed at the final phase of the second stage or the beginning of the third stage. In other words, the use of analytically-processed language continued to be dominant, while the proportion of the holistically-processed language barely increased, so it was difficult to achieve relative balance between the two. In addition, this study proved that the situations of the overuse, underuse, idiosyncratic use and misuse of chunks were universal for L2 learners.To conclude, this study investigated the features of chunks used in CFL learners’ spoken Chinese based on authentic materials, which fulfilled the gap of empirical research on chunks used by learners of Chinese. The results verified the universality of conclusions on the features of chunks used for L2 learners, and it provided data evidence that support the theory of dual-mode system of language process. In addition, this study offered practical recommendations for teaching chunks in CFL.