The topic unit in spontaneous american english: a corpus-based study

Abstract

This thesis is dedicated to the analysis of the topic information unit in spontaneous AmericanEnglish speech as the unit is defined within the Language into Act Theory. In order for thestudy to be possible, a sample of the Santa Barbara Corpus of Spoken American English wasextracted. The sampling strategy adopted and the methodological steps followed in this phasewere carefully chosen seeking to create an American English (AE) minicorpus with the sameoverall structure of the minicorpora of the C-ORAL projects for Italian and BrazilianPortuguese. Accordingly, the sample taken from the Santa Barbara Corpus was submitted toprosodic segmentation into tone units and utterances and to text-to-speech alignment at theutterance level. Following the alignment phase, the AE minicorpus received informationalannotation based on the so-called Informational Patterning Hypothesis, which establishes acorrespondence between the units of the prosodic pattern and those of the information pattern.The topic information unit like all the units of the information pattern is defined withrespect to prosodic, distributional and functional criteria. Regarding the functional aspect, thetopic is defined as the unit which supplies the domain of identification according to which theillocutionary force carried by the comment must be interpreted. After the phase ofinformational tagging, which involved the identification of all information units in the AEminicorpus, the analyses were carried. Both the transcription and the acoustic signal ofutterances containing topic units were set apart for detailed examination. The transcriptionswere saved in spreadsheets and the sound files organized into separate folders. Thespreadsheets were used for the analysis of morphosyntactic features and semantic propertiesof the topics found, as well as for the counts provided in this thesis; the sound files were usedfor the prosodic analysis. In the prosodic analysis, the nuclear portion of the topics wereidentified, and their relevant acoustic parameters were assessed. The topic units were thenclassified with respect to the prosodic form through which they were realized. In this study,certain prosodic features that had never been found in topic units in other languages werefound, which comprises one of the contributions of this thesis to a better understanding of theprosodic codification of the topic unit. Furthermore, the outcomes of this study yielded theproposition of a new account for the possible prosodic realizations of the topic unit. Finally,the AE minicorpus is offered as a contribution to researchers within the Language into ActTheory as well as for the general field of spontaneous speech studies.