Various previous studies have claimed that there are no regional differences within Australian English phonology. The main recent proponent of this view has been Bernard (1969: 62, 1981); see also Hammarström (1980: 42) and Turner (1966). Others, such as Mitchell and Delbridge (1965a, 1965b), while finding a different form of /ou/ in Adelaide, have generally agreed. Conversely, in an ongoing study (Bradley and Bradley 1991) quite a number of regional phonological differences have been identified, one of which, the lexical diffusion involving /a:/ and /æ/ in words such as castle or dance, is a popular regional stereotype and also shows social and stylistic differences. It may also reveal something about the process of lexical diffusion and the phonetic characteristics of the corresponding vowels in southeastern British English at an earlier stage.

Historical background and previous studies

It has been suggested, for example by Samuels (1972), that the eighteenth and early nineteenth century development of Middle English short a to long /a:/ rather than /æ/ before the fricative /f θ s/ (e.g. bath, class) as a reaction to the deconstriction of postvocalic r, preceded the development of /a:/ rather than /æ/ before a cluster of nasal plus obstruent (e.g., dance). The fact that Australian English varies, but tends more to /æ/ in the latter environment, as discussed below, is supportive of this view and may help to order and date the stages of lexical diffusion from /æ/ to /a:/ in England, as the koinéization (Trudgill 1986: 129–46) of Australian English probably took place, according to contemporary reports, in the very early nineteenth century.