Tracking modern human population history from linguistic and cranial phenotype.

Hugo Reyes-Centeno,Katerina Harvati,Gerhard Jäger

doi:10.1038/srep36645

Abstract

Languages and genes arguably follow parallel evolutionary trajectories, descending from a common source and subsequently differentiating. However, although common ancestry is established within language families, it remains controversial whether language preserves a deep historical signal. To address this question, we evaluate the association between linguistic and geographic distances across 265 language families, as well as between linguistic, geographic, and cranial distances among eleven populations from Africa, Asia, and Australia. We take advantage of differential population history signals reflected by human cranial anatomy, where temporal bone shape reliably tracks deep population history and neutral genetic changes, while facial shape is more strongly associated with recent environmental effects. We show that linguistic distances are strongly geographically patterned, even within widely dispersed groups. However, they are correlated predominantly with facial, rather than temporal bone, morphology, suggesting that variation in vocabulary likely tracks relatively recent events and possibly population contact.

Highlights

Explorations on the association between languages and genes indicated that patterns of linguistic diversity paralleled those of genetic diversity
Of the three cranial regions, the highest correlation with language was for the face, followed respectively by the neurocranium and the temporal bone
In order to assess the extent by which linguistic and cranial diversity was patterned by geography, we computed the correlation of L or phenotypic distances (PST) with G

Summary

Introduction

Explorations on the association between languages and genes indicated that patterns of linguistic diversity paralleled those of genetic diversity. While the pattern between geography and phonemes is supported only under strict assumptions of phonemic inventories and accumulation rates[27], the observed loss of phonemic diversity has suggested to some that language traits can be used to reconstruct deep population relationships By this logic, the temporal depth of reconstruction would be at least as far back as the genetic divergence of Khoisan-speaking populations, ~40 ka[28,29], and possibly into the time of the common ancestral population[26]. A positive, statistically significant relationship between land-based geographical distances and biological distances is consistently observed for genetic and skeletal data[20,31], but not for phonemic data[32,33] Such a relationship is expected among languages and dialects from the same language family (i.e. a group of languages whose common descent has been demonstrated conclusively by historical linguists). Language can be considered an ‘extended phenotype’[42], which, like the skeleton, is under influence of non-heritable factors or otherwise not directly regulated by the genotype

Objectives

Methods

Results

Conclusion