Prof. Nicholas Evans (Australian National University)
While linguistics knows an increasing amount about the nature of linguistic diversity, we still have no more than a basic understanding of its causes. Why are there such radical differences, across the globe, in the distribution of linguistic diversity – whether measured in numbers of languages, of language families, or of typological variability? While a partial explanation can be found in the homogenising effects of historically long-term state formation in some parts of the world (e.g. Europe, Korea, Japan) and not in others (e.g. New Guinea), this still leaves unexplained the striking contrasts in diversity found if we compare regions where colonisation is historically recent, e.g. New Guinea vs Australia or Vanuatu vs Samoa. And while various extralinguistic causes have been proposed with some statistical validity (e.g. growing season, ecological risk), at best they furnish distal causes which must logically be mediated by sociocultural differences then played out in different linguistic practices and ideologies.
In other words, to understand the origins of linguistic diversity (macrovariation) and its very different distributions across the world, ultimately we need to investigate differences across speech communities in microvariation – the way small-scale individual differences arise, diffuse, and are semiotically mobilised, across different types of small-scale speech community. In this talk I outline the methods and unfolding findings of a large-scale comparative project which sets out to do just that, focussing on case studies from three high-diversity regions: northern Australia, southern New Guinea and Vanuatu. Our studies of these regions are particularly concerned to examine the role that egalitarian multilingualism and the resultant metalinguistic awareness of linguistic difference play in the ongoing fostering of linguistic diversification at various levels, and with reconciling the need to balance variation that is locally salient (e.g. the variable (ng), relevant in English but not, say, Xhosa) with the need for cross-linguistic commensaribility in levels of typological diversity, e.g. through the typological variables in WALS or GRAMBANK.
If – as our emerging findings suggest – egalitarian multilingualism plays a central role in driving diversification, this makes the documentarian task for the world community of linguists even more urgent than that of studying static structures, since many traditional types of multilingualism are even more fragile than individual languages themselves.