12. Language contact phenomena and genetic classification

Language contact and subsequent restructuring is one of the central themes of the present monograph, above all because African languages have a rich story to tell in this respect. Since most speech communities do not develop “in vitro”, there are probably no “unmixed” languages. Nevertheless – and somewhat paradoxically maybe – the present author holds a rather traditional (“Neogrammarian”) view concerning the genetic classification of languages. The primary reason for holding on to the family tree as a classificatory and phylogenetic model of language relationship relates to the nature of so-called “language mixing” or contact phenomena in general, as well as to important principles of taxonomy (Dimmendaal 1995b; the same position is defended in Greenberg 1999). This position, as a matter of fact, has been defended by different scholars at earlier points in time. Scholars like Hall (1958) already defended the view that creolised languages confirmed the usual applicability of genetic relationships among languages. From the discussion of pidginisation and creolisation processes in the preceding chapter, it should be obvious to the reader that the present author shares this latter view; first, because it is important in phylogenetic classifications to distinguish between inherited and innovated structures, and second, because language typology should not be mixed up with genetic classification, a point also raised again and again by the eminent Africanist Joseph Greenberg. Let us have a closer look, therefore, at two alternatives to the family tree model, one claiming a non-genetic development for specific languages, the other claiming a multi-genetic origin, and review the arguments for and against such terminological usage.