A Study Of The Segmentation And Representation Of Motion Events In Korean

Event segmentation is a process in which the visual inputs presented for the task interact with representations in event schemata,under the conditions of a specific task.Previous studies on event representation have extensively focused on the lexicalization patterns of motion events.The lexicalization typology of motion events is semantic-oriented and fails to fully explain the syntactic-semantic interface.Employing the elicitation experiment,this study conducted two experiments,with 31 native Korean speakers selected as participants.The task of Experiment 1 was to watch 21 short videos representing spontaneous motion events and described what happened in the video.There were two tasks in Experiment 2.Participants firstly described what happened in the 23 videos representing causal-chain motion events,then completed the causality rating task to judge the directness of causation.For the motion descriptions obtained in Experiment 1 and Experiment 2,the sub-event segmentation patterns,morpho-syntactic property of Korean macro-event constructions,and lexicalization patterns were analyzed respectively,so as to explore the typological features of the segmentation and representation of Korean complex motion events.For the data of the causality rating task,a series of mixed-effect models were conducted.The results shows that(1)there is a Goal-bias in Korean linguistic representation;(2)Korean is a verb-framed language,with language-specific features such as serial verb constructions and post-positional case markers.Korean shows different lexicalization patterns in spontaneous motion events and causal-chain motion events.In spontaneous motion events,Korean conflates Motion with Deixis but expresses Path and Manner separately.However,in causal-chain motion events,Korean conflates Motion with Path,but expresses Cause and Manner in separate verbs;(3)Korean shares the typological characteristics of Bohnemeyer et al.’s Type-I languages and Type-II languages.In spontaneous motion events,Korean prefers to encode sub-events with combinations of multiple location-change-denoting verb phrases in a single macro-event expression;whereas,in causal-chain motion events,Korean prefers to use multi-macro-event construction and serial verb construction.The choice of causative construction is closely related to the directness of causation.The results are discussed in terms of language universals,language-specific features,and psychological reality.This study reveals that typological differences are internal differences due to specific event types.Korean is not a single segmentation / lexicalization pattern,but with a pattern as the dominant.The segmentation typology of Korean motion events is mainly type-I language(serial verb language)of Bohnemeyer et al.,but with the characteristics of both type-II and type-III languages(verb frame language);the lexicalization typology of Korean motion events is mainly Talmy’s verb-framed language,but with the characteristics of both satellite-framed languages and equipollent-framed languages.In addition,this study,to a certain extent,complements Bohnemeyer’s typological study,and proves that it is not appropriate to classify Korean into a single type through the typological study of the segmentation of two different types of Korean motion events.