A multi-level analysis of subjecthood diagnostics in Korean

There have been proposed several different diagnostics for identifying the subject in Korean. Among the previous studies, Hong (1991/ 1994) investigated 9 subjecthood diagnostics and classified them into a few groups. This paper is based on previous experimental studies and statistically examines how Hong's classifications affect each subjecthood diagnostics and the acceptability scores of native speakers. Among the 9 subjecthood diagnostics, the experimental data for 6 subjecthood diagnostics were taken and they were statistically examined. This paper organized 6 subjecthood diagnostics into 3 groups following Hong's classifications, and it statistically examined how each group and each diagnostics affected the acceptability scores of native speakers. A multi-level/ hierarchical Generalized Linear Mixed-effects (Regression) Model was adopted to statistically examine the influence of each group of diagnostics. Through the analysis, the following two facts were observed: (i) Each group of classification did NOT play a significant role in the subjecthood diagnostics, and (ii) Honorific Agreement and Plural Copying demonstrated statistically different behaviors from the other 4 subjecthood diagnostics. (Chungnam National University . Hannam University . Seoul National University)