Contact or Inheritance? Criteria for distinguishing internal and external change in genetically related languages
Several prominent scholars have recently doubted whether it is possible to differentiate borrowing from internal change, to the point that in some cases subgrouping is not feasible or is restricted (Dench, 2001; Dixon, 2001). Since a situation of prolonged and intense contact between closely related languages is very common, language contact and its results are a major problem if not a real hazard to historical linguistics. The main practical problem is how to differentiate internal changes, changes motivated by internal processes, from external changes, changes due to language contact, when the structure of the languages is so similar. In other words, how do we know which linguistic form is the source of the change: one of the attested languages, or the mother of both of them? In this paper, I suggest two preliminary criteria to isolate the source language in cases of contact: 1) the existence of intermediary stages, and 2) an even spread of the change across categories. I will show, using test cases from the Semitic language family that these criteria can help us distinguish between internal and external changes.