Formulaic expressions in Korean academic discourse: A corpus-based combinatoric morphemic analysis

Abstract

This study introduces a corpus-based investigation of formulaic expressions in Korean academic discourse, employing a refined morphemic analysis designed for agglutinative properties of Korean. In the corpus analysis, dynamic discourse functions of formulaic sequences in Korean academic prose and formal conversation are explored while determining distinct register-based properties. Over the past thirty years, various corpus-based studies have rigorously examined recurrent formulaic expressions (so-called lexical bundles or multi-word expressions) in English, Spanish, etc. In contrast, there have been few studies in an agglutinative language like Korean with intricate morphosyntactic dependencies. By implementing pre-processing of allomorphs and unnecessary morphological units and lemmatization of predicate endings, the new combinatoric morphemic analysis provides substantial lists of lexico-grammatical patterns with accurate frequency information. This methodological template paves the way for further research into formulaic units in other agglutinative languages like Japanese, Turkish, and beyond. Three types of corpora have been used: a written corpus (2000 academic journal papers), a spoken corpus of formal conversation and a balanced reference corpus (3 million words). The result affirms high productivity and dynamic linguistic functions of Korean formulaic expressions in academic discourse, which indicates their utility as a valuable resource for exploring the process of second language learning and pedagogy.