امید علاف اکبری

The advent of ChatGPT, an all-purpose intelligent Chabot created by Open AI, has created a multitude of prospects and opportunities within the realm of language education. Owing to its exceptional capacity to produce a wide range of textual outputs, ChatGPT responds to queries in a matter of seconds. The present study aimed to examine the appropriateness of ChatGPT-based instruction and flipped language learning instruction (FLLI) in classrooms in terms of enhancing EFL learners’ interactional metadiscourse realization across language proficiency levels. Sixty-three English language learners were divided into four distinct groups: 16 in the advanced and 16 in the intermediate FLLI group, 15 in the advanced, and 16 in the intermediate ChatGPT group. Before initiating the study, all participants were given a pretest that specifically assessed the use of interactional metadiscourse markers (IMMs) in argumentative writing. The four experimental groups were separately exposed to ChatGPT-based instruction and FLLI during eight sessions. Results from a two-way analysis of variance (ANOVA) procedure through SPSS (version 25) revealed that the ChatGPT-based instruction was more effective than FLLI, and both intermediate and advanced language learners following the ChatGPT-based instruction used IMMs in argumentative writing more successfully. Moreover, the participants’ perceptions of ChatGPT and FLLI were analyzed through MAXQDA (version 2020), showing more positive attitudes towards using ChatGPT in language learning in general and IM realization in particular. The findings have the potential to provide advantages for teacher educators in terms of utilizing ChatGPT learning approaches.

Many scholars have long contributed to making the instruction of oral production more effectively. This study compared three task types (information-gap, reasoning-gap, and opinion-gap tasks) and two cognitive styles (field-independence and field-dependence) regarding their effects on English as a foreign language (EFL( learners’ oral production linguistic complexity. The current study was quantitative in research methodology and followed the comparison group design. Initially, 230 Iranian learners were selected using convenience sampling. After the participants sat the proficiency test, the number of participants was reduced to 180 BA students at the university of Mohaghegh Ardabili and Islamic Azad University, Ardabil Branch. They were randomly divided into six experimental groups. Two groups were randomly assigned to field-independent (FI) and field-dependent (FD) information-gap tasks, the next two groups to FI and FD reasoning-gap tasks, and the other two groups to FI and FD opinion-gap task types. Each group consisted of 30 advanced EFL learners and was taught oral production using one of the above-mentioned task types. Michigan test of English Language Proficiency test (Phakiti, 2003) and group embedded figures test (Witkin et al., 1971) were used to determine proficiency level and measure cognitive styles. A two-way ANCOVA (analysis of covariance) procedure was run to examine the data. The results revealed significant differences among task types and cognitive styles, with FD learners and opinion-gap task being the most effective on EFL Learners' oral production linguistic complexity. The interaction between cognitive styles and task types is more likely to account for language learners’ oral performance.

While metadiscourse has been extensively examined across several genres, contexts of publication, disciplines, and languages over the past two decades, researchers have mainly limited themselves to the qualitative checking of candidate metadiscourse markers for the various functions they serve. In the present study, however, we drew on retrospective methods coupled with semi-structured interviews to gain a deeper understanding of metadiscourse features applied linguistics apprentice and professional authors use in their research articles (RAs) in national and international English-medium journals. To achieve this goal, we built on Hyland’s (2019) interpersonal metadiscourse model to analyse RAs in three subsections including introductions, results, and discussion. We ran chi-square tests to examine the RA variations, following the descriptive analysis of the use of metadiscourse markers. A follow-up stimulated recall through semi-structured e-mail interviews was used. We used MAXQDA to analyse the interview data from authors. The results of qualitative and thematic analyses showed that metadiscourse markers play key roles in conveying the writers’ message and intention to the members of discourse communities. The findings of the study suggest raising apprentice writers’ awareness of the way they frame their message in research writing