Volume 8, Issue 3

There is a challenging issue for linguistic typology which involves the relationships which might exist between societal type and aspects of linguistic structure. Linguistic-typological studies have provided us with insights into the range of structures available in human languages, but we do not yet have explanations for why, of all the possible structures available, particular languages select particular structures and not others. A legitimate sociolinguistic viewpoint would be that some social explanations may be available. The sociolinguistic factors suggested as being relevant are language contact versus isolation, and community size and network structure. This paper deals with this thesis from the point of view of Austronesian phonology, with particular reference to Polynesian phoneme inventories.

Languages of the Athapaskan family were often in contact with languages of other families, and speakers were often bilingual (or multilingual). The languages typically have large consonant inventories, and speakers were often bilingual in languages with smaller or larger inventories. These languages present a good laboratory of study for the predictions made by Peter Trudgill concerning sociolinguistic factors governing inventory size. There is relative stability in the size of the stem-initial consonant inventory across the language family independent of the type of contact situation found and the size of the consonant inventories of the languages in contact.

Trudgill has proposed that the size of the phonological inventory for a language correlates with the extent of language contact, with isolated small languages having very small or large inventories, and provides evidence for his claim from the Pacific region. The claim is scrutinised in light of the history and size of the language communities. Language contact is also seen to have contributed to the small and large inventories found in the region.

Evaluating Trudgill's correlation of phoneme inventory size with social factors, this paper highlights the role of phonological structure in the acquisition of phonological contrast, with particular reference to Turkic and Korean vowel inventories. Factors such as social dominance, isolation, and community size are shown not to provide plausible explanations for the loss of primary long vowels in most Turkic languages and their preservation in a few others, nor for the neutralization which long and short vowels currently undergo in Korean. It is suggested that such changes in phoneme inventories can be better understood as a result of phonetic and phonological processes involving the contrastive features and phonological contexts that define the phonemes in question. Processes referring to vowel length in Turkic and Korean are argued to have obscured its contrastive status in the organization of phonological knowledge by the speakers of these languages.

There is no support for Trudgill's thesis. There are languages with exceptionally high numbers of consonants that are spoken by large groups, sometimes by millions as a second language. A pilot survey of contact languages (pidgins, creoles, and mixed languages) does not reveal unusual numbers of phonemes. Non-creole languages that have simplified their verbal infliection due to contact and second language learning appear not to decrease phoneme numbers either. Even most mixed languages that arose among fully bilingual groups fail to show exceptionally high numbers of consonants, even though some may display more complicated structures than their component languages. There is no correlation between language contact and phoneme inventories.

In the target article, Trudgill assumes, based on the inspection of some Austronesian/Polynesian languages, that large community size favours medium-sized phonological inventories, whereas small community size favours either small phonological inventories or large inventories, and he then undertakes to explain these “facts”. A crosslinguistic empirical test, however, reveals conclusively that such assumptions are invalid and therefore Trudgill's explanation is fallacious in explaining a phenomenon that does not exist.