Grammatical Change. Origins, Nature, Outcomes

Abstract

Lexical evidence has played an important role in trying to establish a “Khoisan” language family. With respect to the southern African languages there is indeed a considerable amount of shared vocabulary across all three major established non-Bantu families subsumed under “Khoisan,” viz. Khoe-Kwadi, Kx’a, and Tuu. A historical reevaluation of this phenomenon is presented in a first comparative treatment of body-part vocabulary, including newly collected data. While our research provides support for the above three main lineages (this evidence is not discussed in this paper), it contradicts the view that vocabulary shared across them should also be interpreted in genealogical terms. Such vocabulary can rather largely be explained as the result of different types of language contact, supporting the current dominant view among specialists about the untenability of a “Khoisan” family. From a general perspective, the article argues against superficial unqualified lexical comparison and for a canonical historical-comparative procedure, whereby one reconstructs bottom-up and evaluates at every step whether genealogical relations should be built up further. Although such an approach is deeply entrenched in the traditional method, it is often neglected in many areas of historical language research. We apply it for the first time to the evaluation of the purported “Khoisan” language family and, in addition, venture that contact scenarios should be given more scope in the assessment of historical relations between languages, both in the Kalahari Basin and in general.