Light Verb Constructions in English: Cognitive and Discourse Perspectives

다국어 초록 (Multilingual Abstract)

The term ‘light verbs’ such as have/give/take/do/make in ‘have a swim’, ‘give the show a miss’, ‘take a dive’, ‘do a favor’ and ‘make a decision’ was at first coined by Jespersen(1954). Since then there have been two groups of researches competing in respect to English light verb constructions (LVC, in short). One is to propose that the light verbs which head these constructions do not largely contribute to the meaning of the complex predicate (Jackendoff 1974, Cattell 1984, Baker 1988, Grimshaw and Mester 1988, Culicover and Jackendoff 2001, among others).
The other is to suggest that somehow they put their own meanings to the construction with their contents a little bit lightened (Wierzbicka 1982, Dixon 1991, Brugman 2001, among others). This dissertation is in favor of the latter position, focusing on some problems which have been ignored or overlooked by current studies. In order to deal with the issues, I adopt two perspectives, cognitive and discourse perspective.
Especially, three problems will be highlighted throughout this dissertation. The first one is about allowing a variety of forms of deverbal nouns (DVNs, in short) in ‘true light verb’ (TLVs, in short) constructions, which are defined to preserve the fixed form (‘a + V’), unexpectedly taking other determiners or breaking their frozen status structurally. As DVNs of TLVs have been treated as unstable NPs in that they rarely undergo syntactic operations such as wh-movement, passivization, pronominalization, and occurrence with definite articles, as suggested in Kearns (2002), many spoken data observed in corpora such as COCA, BNC, WebCorp. which threaten their frozen structural status, need to be explained.
The second is about the insufficient analysis of DVN’s own productivity for LVCs in that the appearances of new action verbs in LVC that have generally not been accepted as a member of the DVNs have continually been introduced newly to LVCs in English.
Regarding these two issues, I adopt cognitive mechanisms such as Conceptual Metaphor, Unitarization and Reification. With Conceptual metaphor, it is demonstrated that humans consider even their actions as something possessed and produced, proposing the conceptual metaphors for TLVs AN ACTION IS A POSSESSION, for Vague Action Verbs (VAVs, in short) AN ACTION IS A PRODUCT.
In order to explain the conceptualized path for the various structural forms of DVNs in LVCs, I use other mechanisms such as Unitarization and Reification. In terms of Unitarization, I suggest that people can view an action as a unit by way of considering one action as a representative for the members of its set, which provides a motivation for the process of Unitarization.
On the other hand, for Reification, I adopt Radden and Dirven’s (2007, R & D, in short) proposal that almost all the abstract nouns are regarded as a thing converted from a relation such as verbs or adjectives by way of ‘reification’, and then apply it to the generation of DVNs of LVCs. In contrast with R & D, this study argues that the reification applying to DVNs in LVCs can be classified into four types along the gradient scale with degrees of ‘nouniness’ based on the tests for the grammatical environments preferred for NP: ‘immature reification’ (IR, in short, eg. ‘getting up’): ‘semi-reification’ (SR, in short, eg. ‘a lie-down’), ‘activated reification’ (AR, in short, eg. ‘a walk/walks’) and ‘mature reification’ (MR, in short, eg. ‘marriage’) types. The tests given for dividing types of reification include topicalization, referentiality, co-occurrence with determiners as well as passivization, pronominlization, etc., the latter of which were suggested by Kearns (2002). It should be noted that among these types, SR, AR and MR types are systematically related to DVNs of LVCs under discussion.
The third issue is about indifference toward differences of subtle meanings between LVCs and hea...