102 CHAPTER 6 To summarize, the present study has shown that DUB students generally produce more chunks a er six years in free response tasks and that they attain higher uency and complexity levels than SB students. ese ndings align with other studies comparing DUB with SB students (Gustafsson & Verspoor, 2017; Rousse-Malpat, 2019; Piggott, 2019). In these studies, an increase in chunk coverage also appears to correspond with an increase in uency and complexity levels, which supports the idea that chunks are indicative of pro cient language use and that chunk coverage might even be an excellent additional pro ciency measure as it re ects idiomaticity and authenticity (cf. Verspoor et al., 2012; Hou et al., 2018). Moreover, if the goal of L2 teaching is to su ciently prepare learners to become pro cient and uent second language users with an understanding of how language is used authentically, chunk use should be regarded as an L2 pro ciency measure in its own right. e aim of the present study was to examine the e ect of exposure on the use of chunks and to explore the relationship between chunks and traditional CAF measures. While the study on writing had found that DUB students attained higher CAF levels, this study found that the same DUB students also attained a higher chunk coverage. is paper has argued that a DUB program is thus more e ective in facilitating chunk use, especially longer, lexically based chunks. SB programs with a focus on explicit instruction are equally or more e ective in the shorter and more grammatically based chunks. Secondly, this study has shown that chunks can play an important role in de ning the quality of L2 production alongside traditional CAF measures. e longer, more lexically- based chunks can be considered as “preferred ways of saying things” of native speakers and re ect authentic native-like language. Although learning this kind of chunk is cognitively demanding because of the e ort that is needed to attain mastery, the use of these chunks, once mastered, o ers a reduction of cognitive load because, in the end, these longer chunks are stored and produced automatically as one single linguistic item. ere are two good reasons for chunks to be considered as predictors of L2 pro ciency. First, the use of chunks enhances uency by reducing the cognitive load in L2 production. Second, the use of chunks adds the aspect of authenticity to traditional CAF measures and facilitates the acquisition of a native-like L2 repertoire. Based on the low e ect sizes in the correlation between chunks and complexity measures and chunks and uency measures, we have argued that chunks contribute to complexity and uency, but not directly. And as writing can be complex and uent without being idiomatic, we suggest that chunk coverage should be considered an independent pro ciency measure. is study has a number of limitations. First of all, the sample used for this study is relatively small. A second limitation can be found in the nature of the study. Even though there is no question that the amount of L2 exposure was clearly di erent in the
RkJQdWJsaXNoZXIy MTk4NDMw