English Dialect Alternation
|Submitter Email:||click here to access email|
I'm aware that there is some literature on the use of 'were' instead of 'was' in northern English dialects though I'm not familiar with it at all. I have a query about the basis on which 'were' has been assumed to be the dialect form in use.
I'm a non-native speaker of the dialect(s) concerned, but I do count myself as a ''native listener'' of the Sheffield dialect, having been exposed to it in my mother's and family's speech and having lived in the area for 10 years.
When I first came across this area of research while transcribing recordings of the Survey of English Dialects in Leeds in 1997, it seemed remarkable to me that the forms assumed to be 'were' and 'weren't' had been analysed as such -
for me they always seemed to be 'was' and 'wasn't'. Loss of /z/ is common at least for negative forms in the accent: isn't = in't, doesn't = dun't, hasn't = an't.
The quality of the short vowel in both the dialectal negative and the affirmative is perhaps another indicator of the origin of these forms: it tends to be the short open back rounded CLOTH vowel in British English 'hot', making 'wan't' = 'wasn't' homophonous with 'want' as in 'to want to'. Much more rarely is it the British English NURSE vowel which I take to be a long schwa (cf. many transcriptions of NURSE in Foulkes and Docherty (1999) Urban Voices). I don't think that a form with short schwa occurs.
In other words, I feel that 'were' and 'weren't' may possibly occur occasionally but the majority of reported cases could involve a short CLOTH vowel and therefore be more appropriately derived from 'was', particularly (and perhaps
crucially in a diachronic sense) in the negative, where /z/ loss is usual. The possibility of phonological 'contamination' by 'were' must also be borne in mind. The situation is complicated by the possible occurrence of linking / intrusive /r/ after the 'wa(s)/were' form in some dialects. This often makes it look like the form 'were' _must_ be involved, but a counterexample is in the tag 'worrit' with a clear CLOTH vowel, e.g. 'that wan't a good idea, worrit?'. Perhaps vowel lowering before /r/ could be argued to play a role in this case, but the vowel is short and there is still the CLOTH vowel in the /z/-less 'wan't' form.
These thoughts have just been stirred up again by a word-list of local dialect forms I've been sent by a native speaker of South Yorkshire dialect for transcription into IPA. The forms have been collected by her for use in a project she is carrying out. Interestingly, she has spontaneously used an
orthographic form 'wan't' for the dialect form, not the assumed 'weren't'.
I wondered whether Listmembers familiar with the literature could comment on whether or not the possibility had been addressed that dialectal use of supposedly 'were/weren't' forms actually involved 'wa(s), wa(s)n't' forms, and
whether any instrumental phonetic studies on the vowel quality or some investigation of native speaker intuitions/perceptions/categorisations had in fact been carried out which might justify the analysis as 'were'. Obviously I'd also welcome any comments from native speakers too. I will of course post a summary should the responses warrant it.
Department of Linguistics
University of Cambridge
Sums main page