Background to the project
In England there are increasing reports that the pronunciation of urban speech across the country is becoming more and more like that of popular south-eastern English speech. These changes are illustrated by the well-publicized spread of Estuary English. They have also been revealed by recent research projects into the urban accents of cities such as Newcastle and Derby (carried out by James and Lesley Milroy and Gerry Docherty at Newcastle University), and Milton Keynes, Reading and Hull (by Paul Kerswill and Anne Williams from Reading University).
Scotland has a social and linguistic identity which helps to define the country and its people as distinctly Scottish as opposed to English. This sense of Scottish nationalism has been recently expressed in the 1997 Referendum and the decision for a Scottish Parliament. In these circumstances we would not expect Scottish English to show signs of becoming similar to English English. Thus there was some surprise (and media interest; e.g. Sunday Times, 29.3.98) when a pilot study by Dr Stuart-Smith of a set of recordings of Glaswegian speech collected only months after the Referendum indicated that the pronunciation of certain consonants, e.g. fing for thing, do seem to be similar to southern English English.
The aim of this project was to substantiate these preliminary findings by means of a detailed analysis of Glaswegian consonants in this same data, paying particular attention to variation in pronunciation and how this relates to the social characteristics of speakers (age, gender, background). A statistical analysis was carried out to assess the significance of the findings, and to find out which social groups of speakers share particular accent features.
Significance and originality
This project is the first to consider the pronunciation of Scottish English with explicit reference to changes in the pronunciation of English in England. Our results have also extended current theoretical models of changes in pronunciation, and of language change in general. We are also extending the current methodology for examining pronunciation change, by using the latest instrumental techniques for analysing speech data.
The timing of the research was crucial not only to enable comparison of the results with other recent academic studies of pronunciation in England, but also to provide contemporary information about actual language use in Scotland for professionals working in education and in speech and language therapy. With the majority of phonetics manuals focussing on Southern English speech, clinical speech practitioners working in Glasgow will benefit from insights gained from a new phonetic study of Glaswegian.
Current revisions to advanced Scottish school curricula have created an urgent need for information about the language situation in Scotland. These revised qualifications place a far greater emphasis on the understanding of contemporary language, and language change; one of the three categories in the Advanced Higher Examination syllabus is Variations in contemporary English, in which the student should consider the differences which arise in a speaker's usage which reflect non-linguistic features such as age, gender and social class. (English Unit Specification-Language AH, p.4) The results of this project will therefore be invaluable for teachers and teacher trainers.
In the summer of 1997, a set of high-quality digital recordings was collected by Dr Jane Stuart-Smith and Claire Timmins. Read wordlists and relaxed casual conversations were recorded from 32 speakers, men and women, adults and children, from two areas of the city reflecting broadly middle and working class backgrounds.
The main tasks of the project were:
- A fine-grained analysis of the varying pronunciations of consonants in this speech data, according to the social identity of the speaker (age, gender, background). Following the results of the Newcastle University study, we used both traditional transcription and the latest instrumental techniques for speech processing .
- A statistical analysis of the pronunciation data with respect to social factors. This will consider first the variation within and between social groups, and then the actual clustering of variation and the extent to which this conformed with assumed social categories.