This chapter shifts focus to consider to what extent the behaviours viewed in Chapters 3 and 4 were unique to learners. This is achieved by using a new corpus, the TLC L1 corpus, which is composed of the same exam as in the TLC corpus. However, in this case it is L1 speakers sitting the exam. This allows us to see an overlap between the discourse unit functions selected by L1 speakers undertaking the same tasks as the L2 speakers. The role of micro-structural features, specifically grammatical features, in forming similarities and differences between the two sets of examinees (L1 and L2 speakers) is considered. As part of this, the chapter focuses in on four particular grammatical features – demonstrative determiners, numeral nouns, passives and relative clauses – which seem to link discourse unit to proficiency in the TLC to the extent that they generate differences between discourse unit functions when the TLC and TLC L1 are compared. The chapter also considers, however, the normative nature of the analysis undertaken and notes that individual learners’ performance may vary from the norms examined.

This chapter tests the short-text MDA approach at the micro-structural (turn) level in the TLC. The L2 (examinee) and L1 (examiner) turns are treated separately in an exploration of the discourse functions that are present for each type of speaker. A range of metadata variables are explored to see what effect they have on the use of micro-structural discourse functions. The analysis of learner language finds and discusses six dimensions of functional linguistic variation (L2 communicative functions). When metadata is considered, the findings show variation in learner discourse functions based on the learners’ overall mark and proficiency level in different task types. Functional variation attributable to different L1 backgrounds is also observed. Examiner turns reveal distinct repertoires of discourse functions compared to learners, suggesting the influence of social roles on the discourse of both. Narrative elements are discovered at the micro-structural level. The study sets the stage for further chapters that will explore discourse functions at the macro-structural level, considering their implications for our understanding of discourse analysis and its sensitivity to various factors such as role, proficiency and task.