Identifying Maori English and Pakeha English from Suprasegmental Cues: A Study Based on Speech Resynthesis

Abstract

This thesis investigates the suprasegmental properties of Maori English and Pakeha English, the two main ethnolects of New Zealand English. Firstly, in a Production Experiment the speech of 36 New Zealenders is acoustically analysed. Using the Pairwise Variability Index (PVI) to measure syllabic rhythm, the study reveals that the two ethnic varieties display differing rhythmic patterns, with Maori English being significantly more syllable-timed than Pakeha English. It is also shown that, overall, Maori speakers use a higher percentage of High Rising Terminals than Pakeha speakers. The results relating to pitch suggest that Maori English pitch is becoming higher over time, with young Maori speakers producing a significantly higher mean pitch than young Pakeha speakers. Secondly, a Perception Experiment using 107 listeners is carried out to investigate the role of suprasegmental information in the identification of Maori English and Pakeha English. The ability of listeners to identify the two dialects based on prosodic cues only is tested in seven different speech conditions. The various conditions aim to isolate the precise suprasegmental features participants may use to identify speaker ethnicity. The results reveal that listeners are aware of the differing rhythmic properties of Maori English and Pakeha English, and are capable of tuning into the rhythmic characteristics of a speaker to use it as a cue in dialect identification, with some level of accuracy. The perceptual relevance of other prosodic cues is also discussed and the results indicate that, based on certain stereotypes, Maori English speech is assumed to be low-pitched, monotonous, hesitant and slow in pace. It is also shown that listeners who have had greater exposure to Maori English perform significantly better in a dialect identification task than those who are not integrated into Maori social networks, proving that the linguistic experience of the listener is a key indicator of his or her performance in ethnic dialect identification.