On phonemic diversity and the origin of language in Africa

Abstract

Quentin Atkinson (2011) proposes a theory with the important implication that human language originates and expands from Africa. His theory, however, does not seem fully convincing as it stands because, first, he basically tests only a single logical consequence of his theory (whereas normally theories are tested against a variety of such consequences), and secondly, the body of data he provides may also be reasonably accounted for in different ways. Therefore, additional evidence is required to confirm or disconfirm the theory in a more conclusive manner. Atkinson’s article in Science is centered around the positive correlation between the size of the phonemic inventory of a language and the size of the population speaking that language, in the sense that small populations have fewer phonemes. The intuition behind a similar idea is not new: it was suggested by Trudgill (2004) in the specific form that languages with small populations have either very small or very large inventories, and languages with larger populations favour “medium-sized populations”. In testing Trudgill’s claim, I showed that this specific correlation does not really hold in a database of 428 languages (Pericliev 2004), but a part of this correlation, pertaining to the favouring of small inventories by small-sized populations, was later shown by Hay & Bauer (2007) to hold statistically, including in my dataset. Atkinson replicates Hay & Bauer’s finding, and tries to fit it into a serial founder effect, by analogy to human genetics where it is currently believed that genetic and phenotypic diversity declines with distance from Africa, as predicted by a serial founder effect in which successive population bottlenecks reduce diversity, suggesting an African origin of modern humans. Analogously, Atkinson finds that the number of phonemes used in a global sample of 504 languages, derived from WALS (Haspelmath et al. (eds.) 2008), deAUTHOR’S COPY | AUTORENEXEMPLAR