Multilevel dynamics of language diversity and disparity in Oceania

Abstract

There are more than 7,000 languages on our planet today and they are not evenly distributed over the population. Why might this be? This thesis explores the dynamics of language diversification in Remote Oceania, where some islands have 20 times the number of languages of others. In this thesis I investigate, first, environmental and social factors of language proliferation; second, grammatical and lexical dissimilarity of island groups; third, stability of grammatical features and grammatical conservatism of Oceanic languages; and fourth the particular case of language variation in Samoan.

Remote Oceania consists of Temotu, Vanuatu, New Caledonia, Fiji, Polynesia and Micronesia. The region was settled more recently than the rest of Oceania (3,500 years ago) and entirely by Austronesian speakers. Through studies in archaeology, history and linguistics much is known of migration patterns and societal organisation. The comparatively shallow time depth, relatedness of the languages and established knowledge about the history of this region makes it an ideal base for understanding of language diversification.

There is a large discrepancy between the number of languages per island group in Remote Oceania. There are over 100 languages in Vanuatu and but one in Samoa. What could be the cause of this discrepancy? Scholars have suggested reasons such as societal organisation, environmental factors and settlement time. Statistically modelling languages per island group given environmental, archaeological and societal variables in this thesis shows that the primary factors in the proliferation of languages in the sample are likely to be political complexity, time depth and island area. The impact of political complexity is most probably due to the fact that more hierarchical societies lead to more interactions and cohesion over large distances, which reduces language splitting.

However, counts of languages are not enough when studying language diversification, it is also important to consider linguistic dissimilarity. Do languages of certain island groups stand out as unusually varied in their structure or vocabulary? The results show that no island group stands out as containing unusually high levels of structural disparity, but that Temotu, New Caledonia and, to a lesser extent, Southern Vanuatu have higher levels of lexical divergence than the rest. This finding is consistent with previous literature on "aberrant" Oceanic languages.

This thesis also explores the stability of structural features, and conservatism of Oceanic languages, using computational phylogenetic methods for ancestral state reconstruction. The two computational methods of Maximum Parsimony and Maximum Likelihood methods largely concur with findings in historical linguistics as to the grammar of Oceanic proto-languages. However there are few features which are consistently stable across both methods. The most conservative island group of the Oceanic subgroup is the Bismarck archipelago, which is consistent with established theories of the homeland of the Oceanic subgroup. Among the most structurally progressive are Central Pacific, which runs counter to measurements of the lexical retention rate of the languages there.

Finally, we take a closer look at one language of the region in particular and the language variation there - Samoan. While we do not find regional variation that can cause a split, there is still substantial variation present - but tied to the social sphere. Samoa has not historically been centrally governed, but village district chiefly offices may have been more stable and may have been enough to retard internal diversification and splitting.

In conclusion, change and diversity is conditioned by many local social, historical and environmental conditions - social structure potentially being one of the most powerful. Structure and lexicon, operate under different evolutionary constraints and can tell different stories of the past.