Embedded clause and syntactic complexity in L2 Korean writing: Differences in embedded clause types and learners’ L1

Abstract

Syntactic complexity has long been recognized as a fundamental measure in second language (L2) proficiency assessment, particularly in writing (Foster, 1996; Wolfe-Quintero et al., 1998; Ortega, 2003). However, traditional measures such as T-unit analysis have shown limited sensitivity to Korean’s agglutinative morphology and embedded clause constructions (Kim, 2007; Park & Seo, 2009). Moreover, research on cross-linguistic transfer effects in Korean syntactic complexity remains scarce.This dissertation investigates syntactic complexity development through embedded clause usage in L2 Korean writing, comparing English and Japanese L1 learners across three proficiency levels. The dataset comprises 180 essays from the Korean as a Second Language Corpus. Using statistical analyses, the study examines nominal, adnominal, adverbial, and quotative embedded clauses, focusing on frequency patterns and L1 transfer effects. It also evaluates how reframing embedded clauses through chunk expressions affects developmental interpretations. Under traditional classification, the findings reveal significant proficiency-related increases in embedded clause frequency, particularly in adnominal and adverbial clauses. Japanese learners produced higher frequencies at initial stages than English learners. However, chunk-based reframing revealed systematic deviations from these patterns. For nominal clauses, this approach yielded more gradual and continuous developmental trajectories for both L1 groups, with reduced effect sizes yet sustained statistical significance. Adnominal clauses exhibited the most pronounced change: for English learners, all previously significant differences across proficiency levels disappeared, whereas Japanese learners retained significance only between intermediate and advanced levels. Adverbial clauses remained largely consistent across analytical frameworks, although the chunk-based analysis identified later-stage gains among English learners not detected by the traditional classification. These findings suggest learners’ embedded clause usage aligns with instructional sequencing in Korean L2 curricula. By validating embedded clause types as indicators of proficiency while revealing the methodological sensitivity of outcomes to analytical frameworks, this dissertation advances the understanding of syntactic complexity in Korean L2 writing. The findings further demonstrate that typological relationships can influence embedded clause usage patterns, and that chunk-based analysis can reveal developmental patterns not captured by traditional classifications. Together, these results can offer insights for refining assessment frameworks of syntactic complexity in Korean and underscore the critical role of pedagogical sequencing in its development.