Unusual Syllable Structure in the Kurtjar Language of Australia

The search for linguistic universals has often been frustrating because all it takes to disprove a supposed universal is a single contrary case. In view of this it's not surprising that some linguists have fallen back to positing universal 'tendencies', essentially 'universals' that need not be quite universal and which thus can hardly be disproven. The 'universal condition on preferred syllable structure' once proposed by Hooper (1976: 229-30) is a case in point. The Kurtjar language of Australia has a phonology that is unusual in several ways, including its failure to conform to Hooper's preferred syllable structure, whether or not it can thus be taken to disprove it as a universal 'tendency'. In order to make clear why and how this is so, let's begin by considering proposed universals of syllable structure more generally. Even though one might expect a single case to be enough to disprove a universal, some universals seem to persist regardless. In a phonemic day and age Jakobson (1956: 526) concluded that all language had syllables 2 with initial consonants (C) and ones with final vowels (V), and that the syllable shape CV was thus universal. That phonemic CV syllables are intended is not only implied by Jakobson elsewhere, but also made abundantly clear by Malmberg (1963, p. 129), where the universal is stated unequivocally shortly after a hypothesized primeval language without contrastive vowels, in which phonemic /ptk/ is realized with a phonetic schwa after each consonant, is taken to have syllables consisting solely of the individual phonemic consonants rather than as their phonetic CV manifestation. Sommer (1970) should have disproved the universality of CV syllables when he described certain 'Kunjen' varieties of northern Australia as lacking such syllables, and having only VC(C(C)) syllables instead, i.e. ones with initial vowels and final consonants or clusters. (Whereas Dixon (1970) demonstrated that an Olgolo variety of Kunjen clearly does have final vowels, as Sommer (1969: 31-2) had also noted, this does not detract from Sommer's claim about such varieties of Oykangand.) Subsequently, however, Darden (1971) and Hooper (1976: 199, fn. 3) attempted to maintain the validity of Jakobson's claim by reinterpreting it phonetically: they suggested that some final consonants may be elided or followed by epenthetic vocoids to produce phonetic CV syllables in Kunjen. Kunjen was thus taken to be unexceptional, even though, as Anderson (1974, p. 254) correctly inferred, the way its speakers actually syllabify forms is …