Rhoticity is concerned with the pronunciation of the consonant r. In English, rhotic speakers pronounce all instances of /r/, non-rhotic speakers do not pronounce those instances of /r/ in a syllable coda like ca(r) or ca(r)d (loosely, dropping postvocalic /r/, or simply r-dropping).  All accents of English used to be fully rhotic, pronouncing all instances of /r/ (like brother, cart, hear, hearing).  However, during the middle ages and by the fifteenth century at the latest, East Anglians were pronouncing only the prevocalic instances of /r/, like brothe(r), ca(r)t, hea(r), hearing. This is revealed by variant spellings in documents such as the Paston Letters. East Anglian migrants settling in North America established non-rhotic accents in colonies along the Atlantic coast. In England, during the past 300 years or so, this sound change spread northwards, westwards and southwards from East Anglia. The non-rhotic pronunciation was established in London by 1800, Liverpool and Manchester by 1870, and East Kent by 1940. The last vestige of a dropped /r/ is a schwa-like glide in pronunciations like [mɔə] for more, but this is usually still counted as being non-rhotic. And, of course, the linking r is part of a non-rhotic accent, a usually dropped /r/ reappearing across a morpheme boundary between vowels, for example doo(r) versus door opening, or hea(r) versus hearing. The rhotic boundary to the southwest runs roughly from Gloucester to Dorset today, but is creeping westwards. Lancashire became completely non-rhotic during the twentieth century, apart from the Rossendale area.


The model of language sound change adopted here assumes that a sound change is introduced when the children of a new generation restructure what they hear spoken in their local speech community. This is suitable for the present situation, where recordings are what happen to be available in archives. A model that additionally recognises adults collectively altering their pronunciation during their lifetime would require planned recordings of the same informants repeatedly over a very long period of time, perhaps combined with interviews. One motive for larger numbers to actively modify their accents is to avoid intolerance and scorn from other people who have stigmatized some aspect of pronunciation. Another is to improve their chances of employment by establishments that actively discriminate against some varieties of speech.

The following sets of recordings, available online, were used as data sources (the sets are not necessarily complete, they might be a selection from an original collection, or the current progress in digitization):

  • The Leeds Survey of English Dialects. Directed by Harold Orton (1898-1975) of the University of Leeds. Recorded in the 1950s. Elderly informants from country areas were selected, in order to observe conservative features. Online at the British Library.
  • A collection of British and Commonwealth dialect recordings from soldiers held in German prison camps 1915-18, from the Berliner Lautarchiv, directed by Professors Wilhelm Doegen and Alois Brandl. Online at the British Library. These recordings were cut directly onto 78rpm gramophone disks, the informant shouting into a horn; the sound quality is similar to that of telephone transmission, consonants like /f s h/ being particularly indistinct or even eliminated.
  • The Millenium Memory Bank, recordings by the BBC made in 1998-99. Online at the British Library
  • BBC Voices, recorded by the BBC in 2005. Online at the British Library