London as the origin of change in British dialects? Conflicting processes of innovation and levelling in south-east England
Paul Kerswill, Eivind Torgersen (Lancaster University) and Sue Fox (Queen Mary, London University)
The idea of London as generator of change in British English is an article of faith among sociolinguistically-minded dialectologists. Our paper, reporting Phase I of the first variationist study of London,  challenges this view. We ask:
i. What evidence is there that linguistic innovations start in London and spread out from there?
ii. Does the high degree of multilingualism in London have an impact on London English as a whole? (One-third of London's inhabitants are not L1 speakers of English.)
iii. Is there evidence of a 'multi-ethnic vernacular' among youth? If so, does it impact on mainstream varieties?
The project examines an inner-city and an outer-city borough. They were chosen:
i. to reflect a high proportion of non-'Anglo' inhabitants of immigrant descent (inner London) vs. a low proportion (outer London)
ii. to reflect a contrast between relatively closed networks (inner London: ethnic group and family based) and relatively open networks (outer London: geographical/social mobility and greater prosperity).
Our previous analyses of the vowels of towns in south-east England (Milton Keynes, Reading and Ashford) demonstrated regional convergence - dialect levelling . Our supposition was that these new, levelled (or compromise) vowel qualities originated in and were diffusing from London.
We would expect the teenagers in the new project (N=60, recorded in hour-long ethnographic interviews) to show precisely these levelled qualities, with further progress along phonetic trajectories reflecting the assumed innovatory status of London. However, for the vowels of GOAT, FACE, PRICE, STRUT and FOOT, evidence almost entirely contradicts this expectation. The main patterns are (1) a reversal of 'diphthong shift', which had led to broad diphthongs in London English, and (2) a rejection of the 'levelled', compromise vowels of the south-eastern towns.
These changes are highly correlated with ethnicity, gender and borough. In general, the frequency of these features patterns according to the following three interacting scales:
1. Caribbean/West African > non-Anglo other than Caribbean/West African > Anglo
2. Male > female
3. Inner London > outer London
- to which can be added:
4. Outer London > London periphery (Milton Keynes, Reading, Ashford)
Our conclusion is that there is endogenous change among non-Anglos in inner London, and that this is only affecting other groups insofar as they have network contacts with them. The 'levelled' (perhaps 'Estuary English') variants we previously found in the London periphery do not originate in inner London, but are the product of regional supralocalisation in the South East.