This study reports on a classroom observation of the language produced by intermediate EFL learners in a Korean university, especially in terms of negotiation of meaning. Negotiation of meaning during task interactions makes certain that task participants receive comprehensible input and produce comprehensible output, which have been argued as essential elements for foreign language learning. Task type is also considered important, with those tasks requiring an exchange of information most likely to facilitate the negotiation of meaning. The purposes of this study was to compare successful meaning negotiation in four types of task(information gap, jigsaw, problem solving, and sharing personal experiences tasks) in terms of qualitative meaning negotiation. For this study, I recorded eight different tasks involving twenty-four students, a total of around five hours of learner interaction. This study showed that qualitative meaning negotiation is more important than the mere evidence of meaning negotiation indices for foreign language development. In terms of task type, open tasks such as problem solving and sharing personal experiences tasks may facilitate a higher qualitative negotiation than information gap tasks and jigsaw tasks, especially in intermediate or advanced English classes.

The task-based approach to second or foreign language pedagogy aims to provide learners with a natural context for authentic language use. While learners are performing real-world or pedagogical tasks, they have opportunities not only to get a rich and comprehensible input of real language, but also to produce target language items to exchange meanings. Interaction in doing the tasks is thought to facilitate language learning process. Thus, one of the important things that teachers have to do first for their task-based English classes is to design tasks for target language items reflecting native speakers’ authentic language use. It is expected that learners can communicate with foreigners using the prescriptive target language items outside the classroom. This research attempted to find out if non-native English teachers would be able to make accurate predictions about target language items in terms of language forms and lexical phrases that would naturally occur when English native speakers carried out two types of tasks (closed tasks and open tasks). The results showed that many language items predicted by non-native English teachers did not appear in the recorded data by English native speakers, especially for open tasks. Thus, this research called into question the practice of setting tasks at the end of a PPP cycle (presentation, practice, and production), to allow students to put into use target language items that has previously been practised.

This study investigates how language teachers' pedagogical concerns can find a suitable framework for learner corpora-based error analysis in English classes. It demonstrates the ways in which teachers can benefit from compiling learner corpora and the ways in which they can transfer learner corpus findings to pedagogy. By analyzing the learner corpora and investigating common patterns of learners' errors, teachers can gain insights into learners' interlanguage development and provide feedback in the form of remedial teaching to cater for learners' linguistic needs. Consequently, this will help classroom instruction to be more effectively focused. The method of this study is both quantitative and qualitative in that it starts with a word-based method of analysis and ends with a category-oriented display of results. In addition, some classroom concordancing activities are suggested to present the idea of recycling learners' language output and transforming the output into a new form of input which is assumed to be comprehensible to learners.