Towards a better understanding of contact relations between southern African Khoisan languages of the Kx'a, Tuu and Khoe-Kwadi families

Portuguese Foundation for Science and Technology (FCT): 2022.06259.PTDC
(doi.org/10.54499/2022.06259.PTDC)

Duration: 2023-2026

Principal Investigator: Anne-Maria Fehn

​

​

This project proposes a multidisciplinary approach to the quantification of lexical borrowing between the non-Bantu click-languages of southern African commonly referred to as Khoisan. Uniting expertise from phonetics and phonology, historical linguistics and computational modelling, we will develop an integrated methodological framework in which synchronic patterns are applied to diachronic situations, bridging the gap between language description and historical reconstruction. The quantitative output of our study will allow for comparison with data from human genetics and thereby provide new insights into the linguistic and cultural history of pre-colonial southern Africa.

​

Much like the Indo-European controversy, the origin and mutual relationships between the non-Bantu language families of southern Africa (Kx’a, Tuu, Khoe-Kwadi) have sparked a rich scholarly debate with contributions from multiple disciplines. Languages of all three families have large phoneme inventories featuring click sounds and a shared phonotactic profile, leading to the postulation of Khoisan as a genealogical unit. More recently, this proposal has been rejected on the grounds of lack of evidence produced by the application of the comparative method. Alternatively, similarities between the Khoisan languages of southern Africa are ascribed to intensive contact within a linguistic area roughly delimited by the Kalahari Basin. Evidence from various linguistic domains further suggests that Kx’a and Tuu are typologically closer and constitute the older stratum, while Khoe-Kwadi is a more recent arrival to the area. In consideration of extra-linguistic evidence from archaeology and human genetics, Khoe-Kwadi has been linked to the introduction of pastoralism from eastern Africa, around 2,000BP.

​

While previous studies have identified lexical items from all word classes and semantic domains shared by all three families, they have generally focused on data from a small set of well-documented languages from a limited geographical area. This project will make use of a much wider range of data from both published and unpublished sources, including new data to be collected in the field. Due to the special status of Khoe-Kwadi as a possible immigrant to the area, we will pay particular attention to the attainment of a complete coverage of the family: in addition to dictionary-based data, we will also compile comprehensive wordlists from Kwadi and extinct Khoekhoe varieties, as well as from the near-undocumented Khoe languages from the Eastern Kalahari Basin fringe (Shua, Tshwa).

​

Drawing on our extensive experience in the documentation and historical comparison of Khoisan languages, we will establish cognate sets with a focus on words that extend across all three language families, paying special attention to little studied lexical categories like ideophones and onomatopoeia. This will be done on the basis of regular sound correspondences, the interaction between consonants and vowel phonation, consonant-tone interaction, as well as on the possible effects of features no longer present in the modern languages. Once the cognate sets have been established, we will propose a historically informed set of criteria to assign each item to a source language or family.

​

We will then assess a) the extent of borrowing between families and their subgroups; b) semantic fields involved in the borrowing; c) the impact of borrowing on the phoneme inventories of the languages involved. Finally, we will apply an inference method based on Approximate Bayesian Computation in order to propose a model-based historical scenario underlying the observed amount and semantic scope of shared lexis between Kx’a, Tuu and Khoe-Kwadi. In a final step, we will contextualize our results with publicly available data from population genetics, including populations from all three families.

Based on our findings, we will propose an explicit scenario for the date, amount and intensity of major borrowing events that led to the considerable sharing of lexis between languages of the Khoisan linguistic area in southern Africa.