Geometric morphometric analysis of the 'skeletal vocal tract': A first step in involving osteoarchaeology in the search for a potential 'genetic bias' for language, using two Dutch historical populations

The human capacity for speech and language is one of the characteristics that separates us from other animals. When and how the ability to use complex language emerged has not yet been resolved. Furthermore, it is still not fully understood how the huge variety of different languages that exist on our planet has developed. Important in the production of speech are the cognitive and morphological expressions of underlying genetic makeup, which find their phenotypic expression in combination with complex environmental factors. Current research at the Max Planck Institute for Psycholinguistics is based on a theory of genetic bias that states that subtle differences in genotype, expressed as phenotype, can influence speech production at the individual level. If a population-wide shift in phenotype occurs, subtle differences in speech can propagate and amplify through inter-generational cultural transmission of language. With this theory of genetic bias as a guideline, this study attempts to analyse the morphological variation within and between two Dutch historical skeletal collections: Klaaskinderkerke (13th-17th centuries) and Middenbeemster (AD 1829-1866). A thorough background of the anatomy and physiology of speech production is provided, as well as the relation with the skeletal vocal tract. Using three-dimensional surface scans of the crania from these samples, geometric morphometric analyses are used to reveal, segment and explain shape and size variation in a quantative manner, resulting in highly visual output in the original anatomical context. Multivariate statistics enable an assessment of which combinations of multidimensional shape variables allow a differentiation between the two samples and the sexes. Results show that overall, shape variation in the skeletal vocal tract between Klaaskinderkerke and Middenbeemster was too small to allow for a statistical differentiation between them. Sexual shape dimorphism was also not found to be significant. In constrast, when comparing the four sex-by-sample groups, males from Klaaskinderkerke and females from Middenbeemster did show separation. Size variation allowed a distinction between males and females, which is expected in human sexual dimorphism, and also between the samples, which might be related to climatic circumstances or different lifestyles. For both significant shape and size differences, however, future research will have to further investigate the complex relationship between shape, size, allometry, sex, and external factors. In conclusion, at this time, quantifiable differences in vocal tract shape that could indicate a morphological bias could not be supported. The results obtained in this thesis can function as an addition to this new approach in linguistics. The morphological data that were gathered can be added to a growing normative database for phenotypic variation around the world, to be compared to linguistic variation. In contrast to currently used modelling techniques, osteoarchaeology can contribute a direct assessment of morphological variation through time, aiding in the discovery of potential mechanisms behind the development of the range of languages spoken today. For the samples used here, specifically, further research might reveal if a morphological bias played a role in the development from Middle Dutch into modern Dutch.