100 CHAPTER 6 the hypothesis was that students in the DUB program would produce relatively more chunks than students in the SB program. Another hypothesis was that the use of chunks may be related to sentence complexity and uency. As expected, the DUB students produced relatively more words that could be classi ed as part of chunks than the students in the SB condition (d = 0.73, p<0.05). ese ndings suggest that L2 exposure facilitates L2 chunk recognition and use. And as our previous study showed, learners may discover L2 syntactic and morphological patterns on their own through exposure; however, as the data is not longitudinal, we cannot ascertain that it is speci cally L2 chunk use that enables learners to deduce and learn the patterns (Myles et al., 1998; Perera, 2001; Towell, 2014). DUB and SB a ects types of chunks di erentially. As shown in table 15, the DUB group and the SB group performed equally well in the use of Complements, Particles and Discourse chunks while the SB group performed signi cantly better in the use of Collocations and the DUB learners performed signi cantly better in the use of Structures, Compounds, and Fixed Phrases. ese three types of chunks showed a highly signi cant (p<.001) di erence and very large e ect sizes in favor of the DUB program (with a Cohen’s d of 1.37 for Structures, 1.52 for Compounds, and 0.95 for Fixed Phrases). ese typically longer, lexically based chunks are more demanding in terms of cognitive e ort: language learners need to be exposed to these chunks more o en because there is more to remember. Apparently, repeated exposure to and active use of such sequences helps learners to store the sequences as one item in their memory. e SB approach, with more explicit attention to forms, favored the more ‘grammatical’ chunks like Complements, Particles and Collocations. ese shorter, grammatically based chunks are less demanding in terms of cognitive e ort and even in SB programs, students are frequently exposed to these relatively short chunks, due to the predominant role of grammar and verbs. With regard to Discourse chunks, there are no di erences. is may be explained by the special 30-hour writing program in the nal two years provided to both groups. e program contained lesson series, preparing students to take part in discussions and to write essays, both of which enhanced students’ mastery of discourse chunks, as these chunks are usually considered necessary elements in academic speaking and writing. Our ndings are very much in line with Gustafsson and Verspoor (2017), who compared the use of chunks in a high exposure program with learners enrolled in a Content Language Integrated Learning (CLIL) program with 15 hours of L2 instruction and regular students with 2 hours of L2 instruction. ey also found that the high exposure learners used longer and/or lexically based chunks and that regular students did well on short and/or grammatically based chunks, o en encountered in language instruction. ey suggest that short, more grammatically based chunks seem to contribute more to L2 accuracy than to general uency and authenticity and, due to their relative frequency, are easily mastered in early stages of L2 development, even in
RkJQdWJsaXNoZXIy MTk4NDMw