This review appeared in The Times Higher Education Supplement, 1 December 1978.

Phonetic variations


The Linguistic Atlas of England
edited by Harold Orton, Stewart Sanderson and John Widdowson
Croom Helm, £42.00
ISBN 0 85664 294 0


Harold Orton's death three years ago meant that he did not live to see the publication of the linguistic atlas which represents the culmination of his life's work.

The Survey of English Dialects, which Orton directed, was inevitably a long-term project. It began in the 1950s, when a team of field-workers administered an extensive questionnaire to selected informants in over 300 localities and noted down their responses in phonetic transcription. In the 1960s, the data thus collected were collated and gradually published in the shape of the 12-part Basic Material plus an Introduction. And now at last we have a selection from the same material presented in map form -- the form which enables the reader to take in at one glance a geographical pattern he would otherwise have to piece together laboriously from the Basic Material volumes.

So this atlas reflects a quarter of a century's achievement in English dialectological scholarship. Immense time, labour, and intellectual effort have been expended. Why, then, is one left with such a sense of disappointment?

One reason is that the scope of the atlas is much narrower than its title suggests. The localities investigated were almost all small isolated country villages. So in looking at the isoglosses (linguistic boundaries) dividing up the map of England in this atlas one has to remember that almost all cities and major towns have been ignored (the only exceptions are Leeds, Sheffield, York and London). For example, the map showing whether or not any r-sound is pronounced in farmer quite wrongly includes Liverpool in the r-pronouncing region of rural Lancashire. The Scouse accent might as well not exist. The informants, furthermore, were all elderly agricultural workers [apart from a handful of female 'secondary informants', they were also all men -- comment added 1999, JCW] : no doubt a good source of conservative local dialect, but in age and occupation highly unrepresentative of the population of England as a whole. Perhaps the title should have been The Linguistic Atlas of [Male] Working-class Rural England.

This is presumably the explanation for the survey's failure to record certain well-known phonetic characteristics of particular local accents. The map for butter shows no glottal stops in London. The map for dew shows no occurrence of the yod-less "doo" type in Norfolk. There is no trace anywhere of the Cotswolds kind of pronunciation which makes out sound like the Received Pronunciation of eight (as heard from Ron Hayward [a trade-union leader of the time -- JCW 1999] and caricatured by Kingsley Amis). There is no trace, either of the Bristolian habit which makes area sound like aerial. But this omission has another explanation: the design of the questionnaire, which was so historically-oriented that it did not seek to elicit any of the words which are subject to this l-insertion.

The phonological items in the questionnaire were designed to trace the fate of the various Middle English vowels; and the modern final schwa (spelt -a) did not exist in Middle English. Indeed, the phonetic approach of the survey's scholars is pure nineteenth century: it takes no account of structuralist phonemics, let alone more recent developments in phonological theory.

So one cannot, as a result, discover with any certainty the synchronic vowel-system in each of the localities investigated. One cannot see directly where knows is homophonous with nose and where not, or where cut rhymes with put or lamb with psalm. Looking at a map of pronunciations of ewe, what I should like to know is whether it rhymes locally with new, with know, with now, or with none of these. The ewe map in this atlas is organized instead on a purely phonetic basis, so that places where ewe rhymes with know are separated because of the fact that their vowel is in one place monophthongal, in another diphthongal.

Comparing the maps for words which had Middle English /u:/, one finds that in London house, louse and boughs are shown as having a different diphthong from that of shout, clouds, plough and cow. Is this just a random variation, reflecting what the informants happened to say to the fieldworker on that occasion? It seems improbable that it is allophonic (conditioned by the phonetic environment). It can hardly be phonemic (which would mean that bough and plough did not rhyme for Londoners). The editors offer no discussion of such questions; they seem not to be aware of them.

J.C.Wells


Placed on the web by JCW 1999 April 07

Back to Estuary English page