L2 Korean Phonology: the acquisition of stops by English-and Finnish-speaking adults

Abstract

The purpose of this thesis is to find the reason why attaining nativelike pronunciation is difficult in adult L2A. This thesis attempts to take a purely linguistic approach to find it by hypothising that the acquisition of segmental phonology is more than the physical matter of getting the articulators to move correctly and involves phonological rules and principles. The hypothesis was tested through the L2A of Korean stops, which was investigated in three parts; perception of stop segments in word-initial position, production of stop segments in word-initial position and production of stops involving phonological rules constrained by syntax (i.e. the tensification rule vs. the intervocalic voicing rule).Thirteen British English-speaking adults and fifteen Finnish-speaking adults participated in the experiment. The research subjects were divided into three different groups ('Inexperienced I'， 'Inexperienced n' and 'Experienced') according to the length of exposure to Korean. The subjects in the group of 'Inexperienced ľ were exposed to Korean for one year in their native countries, and the subjects in the group of 'Inexperienced 11' for two years in their native countries. The subjects in the group of 'Experienced' attended a Korean language course for one academic year at least in Korea. Firstly, as for perception of stop segments in word-initial position, both English- and Finnish-speaking learners performed better in discerning geminates from non- geminate segment in general. Especially, the two language groups of learners were native-like in discerning a geminate (AA) from a non-geminate of which the segment is different from the ones in the geminate (B). On the contrary, the Korean stops distinguished by the feature [sg] alone have appeared the most difficult for the L2 learners of Korean to acquire. The English- and Finnish-Speaking learners showed a similar pattern of difficulties in discerning Korean stops regarding the feature [sg]; however, differences between the two language groups were also found in the perception of word-initial Korean stops, which were caused by the absence or presence of geminate in the learners' L1. On the other hand, no progress was made by English- and Finnish-speaking learners in the acquisition of Korean stops in accordance with the length of exposure to Korean. Secondly, the production of word-initial stop segments appeared more successful than the perception of them. The difficulty in producing word-initial stops seemed to be caused by Korean-particular phonological representations rather than controlling the degree of VOT values. As for the developmental aspect, English- and Finnish-speaking learners showed the improvement in the segmental production task according to the length of exposure to Korean unlike in the segmental perception task. Thirdly, the English- and Finnish-speaking I earners performed equally poorly on the tensification rule despite the differences in their L1’s. One reason was that the learners in both language groups were not advanced enough to sense the interaction between the phonological rule and syntactic condition in the Korean grammar. Another reason was orthographic influence. Regardless of the two language groups' similarly poor results in performance on the tensification rule, it was presumed that only Finnish speakers would be able to acquire the Korean-specific rule with the supposition that positive L1 transfer might occur at the even advanced stages of learning. In the light of the findings, it was concluded that the hypothesis of this thesis was supported by results from the experiment. Observing that the L2 learners had far greater difficulty in their production of stops involved in the tensification rule constrained by syntax than in their production of word-initial stops, it is concluded that the difficulty of mastering L2 phonology is due to the complexity of phonological rules applying beyond the component of phonology or across phonological domains in the prosodie hierarchy, some of which provide a means for mapping the syntax to the phonology. Therefore, all the complex phonological rules and principles of a segment must be acquired for the target pronunciation.