Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.784011
Title: Experience and learning in cross-dialect perception : derhoticised /r/ in Glasgow
Author: Lennon, Robert
ISNI:       0000 0004 7969 5801
Awarding Body: University of Glasgow
Current Institution: University of Glasgow
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
It is well known that unfamiliar accents can be difficult to understand. Previous research has investigated the effect of hearing e.g. foreign-accented speech, but relatively little research has been conducted on the effect of hearing an unfamiliar native English accent. This thesis investigates how listeners process fine phonetic detail for a phonological contrast, measuring their perceptual efficiency depending on their level of experience with the working class Glaswegian dialect. In Glasgow, speakers are stereotypically rhotic. However, recent sociophonetic research indicates a trend towards derhoticisation (the phonetic erosion of postvocalic /r/ in working class Glaswegian speech). The potential for misperception exists when listeners hear minimal pairs such as 'hut/hurt', when spoken by working class speakers who realise postvocalic /r/ as an acoustically ambiguous variant, with delayed tongue tip gesture and early tongue body gesture. This makes derhoticised /r/ perceptually very similar to the preceding open back vowel in both 'hut' and 'hurt', leading to difficulty when listeners try to distinguish between /CʌC/ and /CʌrC/ words in general. This thesis begins with a novel dynamic acoustic analysis of the key cues for rhoticity in Glaswegian for such words, demonstrating that minimal pairs such as 'hut/hurt' are acoustically very similar in the Glaswegian working class accent, but remain distinct for middle class speakers. A suite of listening experiments is then described. Experiment 1 was conceived and run as a pilot study to this work, but new analysis of the data assessed the influence of long-term learning, using Signal Detection Theory as a key analytical tool. This showed strong effects of listener familiarity in both sensitivity and response bias. Experiment 2 tested listeners' ability to learn the distinction between 'hut' and 'hurt' word types, with Response Time and Signal Detection analyses finding that listeners least familiar with the accent very quickly matched the response patterns of listeners with an intermediate level of experience. However, neither listener group was able to match the performance of listeners most familiar with the accent, who were native to Glasgow, suggesting that acoustic phonetic detail plays an important role in perception, with interesting interactions with listener experience. Finally, for Experiment 3 the overarching purpose shifts from offline to online perception. The results of the acoustic analysis showed that dynamics are key, so Experiment 3, a Mouse Tracking experiment, allows for the measurement of dynamic perceptual responses - in a listening context which is more difficult - for the most experienced listeners. It yielded results in terms of Response Time, and two sets of measures capturing cursor trajectories, Area Under the Curve, and Discrete Cosine Transformation. Taken together, the results of these analyses reveal that even for the most experienced listeners (Glaswegians), the phonetically ambiguous tokens present perceptual challenges when hearing working class Glaswegian 'hut' and 'hurt' words, and also demonstrated that challenging listening conditions lead to processing costs, even for the 'easiest' stimuli; i.e. when hearing talkers and accents randomised together. This thesis examines a single difficult phonological contrast, with the simplicity of the linguistic scope affording an extremely in-depth analysis. Not only did this provide a clear insight into the perception of the contrast itself, but the depth of analysis allows for a more sophisticated discussion of the results, potentially speaking to wider theoretical standpoints. The results have implications for theories of speech perception, as they may be explained by some general principles which underlie exemplar theories and Bayesian inference. They also constitute valuable acoustic and perceptual contributions to the ongoing research into the complex and changing nature of postvocalic /r/ in Scotland.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.784011  DOI:
Keywords: P Philology. Linguistics
Share: