ABSTRACT
Australian languages typically have small vowel inventories and very unusual consonant systems. They have no close vowels, no voicing contrast, and no manner-of-articulation contrasts among the obstruents but they have an unusually large number of places of articulation. Observation of spontaneous speech and acoustic measurements of laboratory speech indicate an overwhelming imperative to preserve place of articulation distinctions, especially in medial consonants. Controlled experiments on the articulator)’ dynamics and micro-prosody of these languages indicate that many of the durational contrasts and nearly all of the spectral contrasts are found in the medial consonants rather than in the vowels. This phonetic enhancement of the acoustic and perceptual salience of intervocalic consonants is manifested synchronically at the phonetic level in most currently spoken Australian languages. It has become phonologised to varying degrees in a number of these languages, whose phonological structure is now arguably VC(V), rather than the universally accepted CV(C). However, languages that have reached this stage may not retain this structure, but show signs of moving towards a CV(C) pattern once more. Thus this may be a cyclical process, possibly driven by a tension between a universal linguistic preference for CV(C) structure in the human brain on the one hand and the unique phonetic characteristics of Australian languages on the other. These characteristics appear to be motivated by the “place-of-articulation imperative”, which strongly favours the VC(V) structure as a means of maximising the perceptual distinctions between consonants. I conclude with some speculations as to what might be the motivation for the place-of-articulation imperative.
