Areal features and linguistic areas: contact-induced change and geographical typology

[Extract] What are the reasons for similarities between languages? Linguistic categories can be similar because they are universal-for instance, every language has some way of asking a question or framing a command.\ud Occasionally, two languages share a form and meaning combination by pure coincidence. Goemai (Angas-Goemai subgroup of Chadic, Afroasiatic: Birgit Hellwig, p.c.) and Manambu, a Ndu language of the Sepik area, happen to use\ud a:s for 'dog'. The word for 'tail' in Kwoma, from the Sepik area, is kinyi 'tail', and that in Manambu is gǝñ-a spurious similarity with Yidiñ (Australian) gini 'penis' is striking. And numerous languages of the world have a negator ma:, just like Manambu, or unrelated Tariana, an Arawak language from Amazonia.\ud Distribution of typological features can vary in significance. Marked nominative systems are concentrated in eastern Africa. In the border region of Uganda, Kenya, Ethiopia and Sudan, nearly all the languages-both\ud Nilo-Saharan and Afroasiatic-with case-marking display this pattern. It is extremely likely that this shared pattern results from diffusion and numerous instances oflanguage contact, and may be considered an areal property (König\ud 2008a, 2008b). König hypothesizes that the direction of spread of marked nominative-patterns was predominantly from Afroasiatic languages into Nilo-Saharan languages. That Berber languages spoken to the north of the cluster\ud of marked nominative languages also have a marked nominative system is significant: the pervasiveness of this property in Berber points to its antiquity in Afroasiatic