Soyeon Yoon, Rice University
|The semantics of connective morpheme eta in Korean serial verb construction|
In Korean Serial Verb Constructions (SVC), two verbs are connected by a connective morpheme e, -ko, or eta (V1-e/-ko/-eta V2). There is research on the constructions connected by e and ko (S. Lee 1993, J. Lee 2001); however, the -eta construction has not been systematically studied. In this study, I will discuss the semantic properties of connective morpheme eta from a cognitive perspective.
The morpheme eta connects two verbs comprising an SVC. I argue that there are two types of V1-eta V2 constructions. One is the type where V1 and V2 share both subject and object (-eta SVC1). This type is productive in that it allows unlimited number of sets of V1s and V2s. In this case the full lexical meaning of V1 is preserved and provides information about the manner in which V2 is accomplished (see example 1). The other is the type where V1 does not assign any arguments while V2 assigns the arguments of the sentence (-eta SVC2). This type is not a productive construction in that the possible V1s are limited to a few verbs which specifically denotes Path such as oli-lift, nayli- lower, tuly- get in, and so on. In this type, the V1 denotes the Path of V2 (see example 2). In addition, V2 takes such verbs as po- see and peli- throw away which are likely to be used as auxiliaries denoting the aspect of V1 if they occur in another SVC (see example 3).
I attempt to present a semantic analysis for the morpheme eta in the two types of V1-eta V2 construction. Basically, -eta profiles V2 while maintaining the most salient property of V1 and this morpheme conveys the influence of this property to the action of V2. Therefore, in eta SVC1 the action of V1 can be interpreted as Manner of V2. On the other hand, in eta SVC2, V2 is profiled due to the morpheme eta and thus its meaning as a full verb is revealed, not being employed as an auxiliary. The morpheme -eta also represents conceptual separation which leads to the temporal or local separation between V1 and V2. For this reason, -eta SVC1 denotes that the action of V1 happens in a certain location and the action of V2 happens in another location later (compare example 1 with example 4). However, in eta SVC2, only the salient meaning of Path is maintained throughout the action of V2, the temporal and local separation is not applicable to this type
(1) na-nun wumwul-eyse mwul-ul tt-eta masy-ess-ta <-eta SVC1>
I-TOP well-at water-ACC scoop-ETA drink-PST-DEC
I drank the water (somewhere else) by scooping it from the well.
(2) na-nun kep-ul nayly-eta po-ass-ta <-eta SVC2>
I-TOP cup -ACC lower-ETA see-PST-DEC
I looked down at the cup.
(3) na-nun kep-ul nayly-e po-ass-ta <-e SVC>
I-TOP cup -ACC lower-E see-PST-DEC
I attempted to lower the cup.
(4) na-nun wumwul-eyse mwul-ul tt-e masy-ess-ta <-e SVC>
I-TOP well-at water-ACC scoop-E drink-PST-DEC
I drank the water (at the well) by scooping it from the well.
Lee, J.-H. 2001. A Cognitive Approach to Connective Particles e and ko: Conceptual Unity and Conceptual Separation in Korean Motion Verbs. Japanese / Korean Linguistics 9. 225-238.
Lee, S. 1993. The Syntax of Serialization in Korean. Japanese / Korean Linguistics 2. 447-463.