Complexity in spoken language: A corpus-based learning approach

The present study set out to explore the change in the complexity of L2 spoken language over time as a result of corpus-based learning activities in the EFL context. In that regard, the study adopted a mixed research design counting on both quantitative data, which was obtained through a pre-test/post-test experimental study, and qualitative data, which was gathered through questionnaires and semi-structured interviews to delve more into the process in which the learners gained a hands-on experience with concordance lines. The participants of the study were 56 EFL learners studying at an English preparatory school at the tertiary level in a fully online learning environment. The spoken data from the pre-test and the post-test was transcribed and analyzed for lexical, phraseological, and morphological complexity indices. The data from the questionnaires and the interviews enabled a closer examination of the participants' perceptions on corpus use. The findings revealed that employing a data-driven learning approach in speaking instruction aided the improvement of lexical and phraseological complexity of oral production. However, direct corpus use did not have an impact on morphological complexity regarding verbs and nouns. Furthermore, the study emphasized the analysis of multi-word units as well as single lexical units in characterizing L2 oral performance development. The study also concluded that while L2 learners benefited from direct corpus use, they found the corpus resources inefficient in practicing vocabulary for speaking due to the lack of audio content, which calls for redesigning the corpus tools to better serve L2 teaching / learning purposes. Keywords: data-driven learning, corpus-based learning, spoken language complexity, phraseological complexity, lexical complexity, morphological complexity