Differential diagnoses of language change in Ngoreme, a Bantu language of Tanzania

Ngoreme (Bantu JE401) is spoken in the Mara Region of northwest Tanzania, and presents an interesting case study in the attempt to disentangle contact from genetic inheritance. Ngoreme is located geographically in between two major subgroups in the region, 'North Mara' (NM) and 'South Mara > Western Serengeti' (SM > WS) (Roth 2018; Schoenbrun 1997; Walker 2013). Ngoreme shares a series of typologically unusual features with each of these subgroups, e.g. asymmetric vowel distribution and auxiliary inversion with NM, height dissimilating prefixes and final vowel deletion with SM > WS. We entertain two main possibilities for the genetic classification of Ngoreme: (1) Ngoreme belongs to both subgroups, as a "cladistic hinge" (Pelkey 2015), and (2) Ngoreme does not belong to both subgroups; therefore, one set of features are shared innovations, while the other set of features are due to contact. We argue that the current evidence better supports the conclusion that Ngoreme is a SM > WS language that borrowed a host of lexical items, as well as an assortment of other linguistic features. In making these arguments, we explore some further implications for the study of language change in general.