UC Santa Barbara

This dissertation proposes a protocol for the study of grammar in interaction. Starting from the collection phase and transcription phase of naturally occurring discourse data, I describe which features should these data possess and how they should be transcribed for future studies. I then take data that were produced following the methodologies discussed and perform a study of reported discourse constructions in Kazakh discourse.

The first study provides an analysis of criteria that are currently employed by corpus designers for the selection of speech events to be featured in a corpus. Through the analysis of different speech events collected in Kazakhstan, Italy, California, and Oaxaca, I demonstrate how current criteria of data selection fall short of representing language use in life as they rely on rigid interpretation of linguistic structure usage and genre conventions. I propose as alternative approach that researchers evaluate speech events based on their organicity in the participants’ lives, and focus only on those events that are consequential on speakers’ lives. Implementing this strategy respects the interactional dynamics of speech events.

The second study addresses the question of whether intonation units are a language universal and of whether they can be reliably identified and annotated by untrained listeners regardless of their level of knowledge of a language. I test 345 untrained listeners (native speakers and non) on their perception of intonation unit boundaries in naturally occurring Kazakh conversation. The results from this study suggest the existence of an effect active for every group of participants. The higher is the inter-expert agreement over the placement of a boundary, the more likely it is that untrained participants will place a boundary in the same position. This result suggests that intonation unit perception happens on a scale, with a tendency on the participants’ side to agree with experts in clear-cut cases.

The third study approaches the study of reported discourse in Kazakh through the analysis of predicate utterance constructions in naturally occurring discourse. First, I provide an overview of the frequencies and distributions of different reported discourse constructions in Kazakh discourse. Then, I identify preferred discourse profiles for utterance predicates and build on this finding to demonstrate how discourse configuration of the converbal form of the utterance predicate ‘say’ dep can partially explain the grammaticalization of this element into a complementizer.

Each study is designed to address one step of a research program that seeks to find explanations for grammatical patterns in discourse. From data collection and handling, to data transcription, to the case study on reported discourse, my dissertation presents the theoretical and methodological passages that underlie the study of language in use.