Second language acquisition of Korean evidentiality in expressions of psychological state of mind

Abstract

This study examines the second language acquisition of the evidentiality requirement in the Korean psychological state of mind expressions. When an experiencer of the psychological state of mind is different from the speaker, Korean language requires an evidential expression to the psychological predicate so that the speaker indicates the source of information of someone else's inner state of mind. Three types of variables were tested in a series of experiments: 1) Learner variables included whether their native language has the similar evidentiality requirement (Japanese versus English), proficiency (High versus Low). 2) Linguistic variables were evidentiality agreement (1stand 3rd person subjects were used with or without evidential marking), and the types of psychological predicates (sensory adjectives versus desiderative adjectival phrase). 3) Lastly, one experiment employed two different contexts where the information of psychological state of mind was obtained by the speaker (See versus hearsay condition). Experiment 1 had participants watching a short video where one character had a contact with another character and then made a comment including psychological state of mind. Participants were asked to choose which one had experienced the emotion based on the comment. Experiment 2 was Grammaticality Judgment Test based on listening to a sentence in a context, where the experiencer is clearly either the subject or someone else. Experiment 3 was an open-ended picture-description production task. Participants were given a series of pictures of people showing various emotions and encouraged to report their feelings to a third party. The results showed all the variables had significant effects on learners' performance. Especially, the Japanese learners whose native language had the similar evidentiality requirement significantly outperformed the English-speaking learners whose native language without such constraints,regardless of their proficiency levels. The GJT showed even larger gap between two native speaker groups, whereas there was no difference in accuracy between low-proficient English speakers and highly proficient English speakers. Despite the English-speaking participants' overall inferior performance in receptive skill tasks, these learners showed surprising evidentiality strategies in the production tasks. English speakers produced much more, attempted to employ valid evidentiality strategies.