Use of a Thematic Template in L2 Korean Relative Clause Comprehension

Abstract

This study aims to capture a comprehensive picture of the heuristics internalized and exhibited by L2 Korean learners, across case-marked and non-case-marked environments. It further examines how perceptual processes—such as segmentation and grouping—underlie these heuristics and seeks empirical evidence for the cognitive costs associated with the functional relabeling of structurally specified NVNV strings. This dissertation comprises three studies investigating the formation and use of NVN and NVNV templates in the comprehension of Korean relative clauses by L2 learners. Together, these studies offer insights into the heuristics employed by L2 users when processing Korean relative clauses. The functional labeling of bare NVN strings by English-speaking learners was not initiated based on the heuristics rooted in the canonical order of their L1 English (i.e., NVN as SVO). Instead, they directly adopted the way of labeling the word sequence of the target structure (i.e., [RC NV]N as [OV]S), similar to how native Korean speakers do. However, L2 users made errors primarily due to the difference in head-directionality between L1 English and L2 Korean, resulting in structurally constrained internal relations in Korean. In typical RC sentences, case markers are employed, requiring the parser to accommodate case-informed NVNV sequences during processing. Consequently, the perceptually internalized template of bare NVNVs (i.e., a mental model) is guided to adjust to one of the four possible variations of RC NVNV strings that include nominative and accusative markers. I argue that this adjustment process enables appropriate interpretations and can be understood as a functional relabeling phenomenon in L2 Korean learning.