Genetic influences on tone? New experimental evidence and its consequences for linguistics.

Abstract

In 2007, we used a cross-linguistic statistical analysis to show that there seems to be a negative correlation between the population frequency of the “derived” alleles of two genes involved in brain growth and development, ASPM and Microcephalin (or MCHP1), and the use of tone in the language(s) spoken by the population. Despite our best attempts at disentangling the confounding effects of contact and genealogical inheritance, and the fact that these correlations stood out among millions of similar correlations between multiple genetic markers and linguistic features, this was, at best, a statistical association and, at worst, a false positive. However, new experimental evidence has recently been published independently of us, showing that the negative correlation between the “derived” allele of ASPM and tone perception and/or processing is present at the level of inter-individual differences among more than 400 native speakers of Cantonese, providing strong support for our initial findings (on the other hand, there is no signal for MCPH1). Here we contextualize and summarize these experimental findings, we address some remaining puzzles (including MCHP1 and the absence of tone in Australia), and we discuss the implications of these results for the language sciences.

Tone and genes: New cross-linguistic data and methods support the weak negative effect of the “derived” allele of ASPM on tone, but not of Microcephalin

“…Nevertheless, the work presented here shows the main limiting factor remains the availability of good quality genetic data with good geographic and linguistic coverage (hopefully, such data will become available from more populations across the globe, and maybe even from past groups using ancient DNA techniques), coupled with appropriate methods of analysis. For example, there are only 10 data points from the Americas and 10 from Papunesia, and none from Australia, the latter being a particularly interesting case [54]. If, indeed, ASPM-D has a weak negative effect on tone, and if ASPM-D has a low frequency among the Aboriginal Australian populations (as could be reasonably expected given the age of the allele and the apparently long genetic isolation of Australia [40,100]), then we would expect to find at least some tone languages among its Aboriginal languages, but this is clearly not the case [34,35].…”

“…If, indeed, ASPM-D has a weak negative effect on tone, and if ASPM-D has a low frequency among the Aboriginal Australian populations (as could be reasonably expected given the age of the allele and the apparently long genetic isolation of Australia [40,100]), then we would expect to find at least some tone languages among its Aboriginal languages, but this is clearly not the case [34,35]. However, there are at least two solutions to this (potential) paradox [54]: first, the small negative bias of ASPM-D on tone is inherently probabilistic and, moreover, does not place constraints on language change and evolution at (very) low frequencies of this allele, so that it is not entirely inconceivable that accidents of history and linguistic expansions and extinctions have resulted in a linguistic Australian landscape that does not include tone at this moment. The second is based on the suggested longterm effects of a high incidence of Chronic Otits Media (COM) among the Aborigine populations on the phonetics and phonology of Australian languages [101], in particular the loss of sensitivity in the low frequency range which, naturally, should explain the absence of tone distinctions (just as it explains a lack of voicing contrasts).…”

“…More importantly, given the controversies the original papers generated (including accusations of implicit racism) [49], subsequent work did not find any influence of these "derived" alleles on cognition, brain size or any other such phenotypes [50,51] (with some debatable exceptions; [52,53]). Therefore, while the crucial roles played by these two genes in the evolution and ontogeny of the human brain are clear, with very probable episodes of positive selection in our lineage [43][44][45], it is currently unclear what phenotypic effects their "derived" alleles may have [26,54] and odds are that the recent and current evolution of these particular alleles is mainly driven by neutral demographic processes and not by selective pressures [47,48,54]. Each circle on the maps represents a unique sample, and its colour represents the frequency (%) of the "derived" allele in that sample.…”