UCLA

Analyzing naturally-occurring conversational video data using the frameworks of interactional linguistics and discourse analysis, this dissertation addresses some of the issues related to the study of reported speech. For the purposes of this analysis, I define reported speech as the ways in which reporting speakers negotiate between the referential content of a reported utterance and the embodied form of speaking that displays their stance towards the reported character and his or her talk. The study aims to contribute to our common understanding of reported speech constructions by describing general, formal features of the constructions through a corpus-based analysis of video data, and analyzing Korean speakers' choices between the proximal deictic reporting verb ile- `go like this' and the distal verb kule- `say like that'.

Corpus-based analyses of the transcripts of the video data show that speakers use reported speech constructions during spoken interaction very differently than they do in written discourse. Due to the characteristics of spontaneous conversation, in which interlocutors constantly monitor each other's action, reporting speakers choose to use one mode of reported speech (e.g., direct speech) above another (e.g., indirect speech) according to the unfolding social situation. They frame reported speech using one reporting verb above the other, and quote the utterance of a single participant instead of a speech exchange where two or more participants are involved. Thus, the analysis of Korean conversational data suggests that speakers' grammatical construction of reported speech is locally and interactively organized rather than static and predetermined.

A detailed analysis of the interactional and sequential context in which the deictic reporting verbs are used shows that the selection of one deictic form over the other - a choice that allows speakers to negotiate between the referential meaning of the reported utterance and the embodied form (i.e. stance) - is made along with observable degrees of animation done by the reporting speaker. This shows that reporting speakers do not simply produce a stream of reported utterances, but also individuate the referent fused into it. The proximal verb ile- `go like this' is used by reporting speakers to communicate with the addressees the embodied form of speaking being animated perceptually visible or immediate and temporally proximal `inside the boundaries of the current interactive field.' The distal verb kule- `say like that' is selected by reporting speakers to individuate and communicate with the addressees the referential meaning that is animated as occurring in the story world, being animated `outside the boundaries of the current interactive field. The predominant occurrence of the deictic reporting verbs in casual conversation shows, first, that the components fused into the reported utterance, such as the referential meaning and the embodied form, are not equally communicated among the interlocutors. Instead, one is foregrounded while the other is backgrounded (or vice versa) according to the unfolding talk in interaction. Second, the choice to employ one deictic reporting verb over the other is an interactional resource for both speakers and addressees to achieve this type of communicative goal.

The findings from this study of the interactional functions of the deictic reporting verbs in Korean conversation can shed light on how speakers in other languages make use of the referential meaning of reported speech and the embodied form of speaking that displays stance in communication.