Wim Gombert

136 CHAPTER 8 (1) Maximal target language exposure, meaningful use within a meaningful context and practice of whole phrases, chunks, and sentences (FUMMs) should be considered as the main conditions for success in all language skills. (2) Much research from a DUB perspective has shown that trial and error is part and parcel of L2 development, so teachers should refrain from providing corrective feedback for all errors, as this may also cause anxiety with L2 learners. An L2 program with abundant L2 exposure (through the teacher, through peers and through texts and videos) is likely to provide enough correct and repeated input to adequately facilitate language development. (3) An L2 teaching program can make use of (traditional) coursebooks with some adaptations: ey should be supplementedwith rich, meaningful and comprehensible oral input, preferably multi-modal, with a high degree of repetition built in over time. Assignments should engage students a ectively and cognitively and provide abundant opportunities for students to actively use the L2 for communication. (4) Attention to form or forms should always be combined with attention to meaning. e acquisition of most morpho-syntactic features of an L2 can reach a su cient level naturally and implicitly. A focus on accuracy is likely to impede language acquisition because it may enhance anxiety with L2 learners. (If the method includes explicit attention to speci c grammar and thematic vocabulary, the pages can be skipped as in Piggott’s (2019) experiment). LIMITATIONS AND FUTURE DIRECTIONS Although the results of this study show that a DUB approach is more e ective than a more traditional SB approach in developing listening and speaking skills and equally e ective in reading and writing skills, several limitations need to be considered. As classrooms are dynamic environments involving numerous psychological and social processes, long-term classroom research usually involves many variables which cannot be controlled for. Henceforth, experimental control, which is a prerequisite for generalization of the ndings, is impossible with this kind of research. Secondly, although this study is uniquely longitudinal in the sense that it investigates the e ects of a 6-year program, similar studies are needed with similar results to overcome the scienti c problem reported above. Both the nature of this research and the lack of replication should limit the scope of this study and caution the reader to generalize any of these ndings. However, although generalization of this study’s ndings might not be possible due to the lack of experimental control, this study’s ecological validity is extremely high, as