Corpus-based Arabic language pedagogy for Turkish learners: A bilingual Arabic–Turkish model for data-driven learning

Abstract

Linguistics has experienced a major transition over the past few decades, shifting from predominantly structural approaches to data-driven approches based on authentic language use. This development has strengthened the role of computational linguistics within traditional linguistic disciplines. Despite the growing global interest in teaching Arabic as a Foreign Language (AFL), instructional materials continue to rely heavily on artificial language samples that do not accurately reflect real communicative practices. This study investigates the pedagogical benefits of corpus-based text sets for Turkish learners of Arabic. The research adopts a mixed-methods approach and employs the following procedures: first, it identifies and analyzes the pedagogical benefits of corpus-based instruction; second, it compares learning outcomes using a bilingual Arabic–Turkish corpus containing approximately 100,000 words drawn from literary texts representing classical, modern, and contemporary periods. AntConc software, which enables comprehensive corpus-analysis through concordances and collocation tools, is employed to examine frequency distributions, collocational patterns, n-grams, and contextual concordances, thereby providing a systematic and replicable corpus-informed analytical framework. The findings demonstrate, through illustrative examples, how discourse is semantically and pragmatically structured, as evidenced by common collocations associated with the word “president” in political discourse and the diverse cultural and social contexts surrounding the word “Arabic.” The results indicate that corpus-based text sets support vocabulary development, enhance grammatical competence, and strengthen communicative abilities by exposing learners to authentic linguistic input rather than decontextualized language examples. Furthermore, the study contributes to the integration of corpus linguistics into AFL teaching by proposing a data-driven pedagogical framework and emphasizing the need for corpus-informed teacher training. It also highlights the value of interdisciplinary collaboration between computational linguistics and language education, positioning corpus-based instruction as a scalable, empirically grounded, and pedagogically robust approach that responds to current developments in applied linguistics.