Received Pronunciation (RP) has been widely described linguistically (Wells 1982, 1991, 1997), although little sociolinguistic research has been carried out on it (Fabricius 2000). Over the last few years, a new trend has been observed in young RP speakers to incorporate non-standard features in their accent, such as T-glottalling (Fabricius 2000).

This quantitative sociophonetic study analyses to what extent T-glottalling is present in the speech of young RP speakers and which are the linguistic and social constraints that affect its variability. The data is based on sociolinguistic interviews of 20 teenagers, aged between 13 and 17, from three different types of schools in the South of England: a major boarding public school, a non-boarding private school and an outstanding rated comprehensive school in a wealthy rural area. This data is compared to 15 older speakers, aged 27, who are alumni of the schools under study. The quantitative data is analysed through multivariate analysis.

Results show that t-glottalling in RP is a well-established feature in word-final contexts and change is in progress in the word-final pre-pausal and pre-vocalic (back vowels) environments. Language change in RP in word-final contexts is being influenced by a set of commonly occurring phrases in informal speech, which contain high frequency monosyllabic words. However, in word-medial contexts, RP speakers remain conservative and change is not visible. (t) Tokens in word-medial contexts mostly belong to low frequency words, therefore possibly contributing to the slow progression of t-glottalling in these environments. Type of school and age are crucial factors in explaining the variability of the glottal stop in RP, with teenage speakers belonging to the most elitist private boarding schools considerably resisting the adoption of t-glottalling and with teenage speakers from the private non-boarding and comprehensive schools leading the changes in word-final contexts.