Friday, February 24, 2017

The Origin of Mid Vowels in Siwi

How does a language with a relatively small vowel system react to pressure from a language with a larger one?

Most northern Berber varieties have a simple four-vowel system: tense /a/, /i/, /u/, vs. lax schwa (/ə/, written e in the official orthography), the latter being mostly predictable and limited to closed syllables. In the eastern and southern Sahara, however, we tend to find slightly larger vowel systems, and it looks very much as though proto-Berber had a rather asymmetrical six-vowel system, close to modern Tuareg but missing /o/: it had tense /a/, /e/, /i/, /u/ vs. lax /ɐ/, /ə/.

Siwi Berber, in western Egypt, has a more symmetrical six-vowel system: tense /a/, /e/, /i/, /o/, /u/ vs. lax /ə/. All of these vowels occur in inherited vocabulary as well as in Arabic loanwords. It is obvious by inspection that, in almost all contexts, *ɐ merged into /ə/. But the distribution of /e/ shows little connection with that of *e: in fact, most instances of proto-Berber *e correspond to Siwi /i/. And the origin of /o/ is not immediately clear at all. How did this happen?

My latest article - written together with Marijn van Putten - proposes some answers. It turns out that proto-Berber */e/ was retained in Siwi only before word-final /n/. Most instances of /e/ and /o/ are found in Arabic loanwords. Within inherited vocabulary, almost all instances of /e/ - and all instances of /o/ - are phonetically conditioned innovations, arising from at least three distinct regular sound changes and one sporadic one. The net effect of this "conspiracy" of sound changes is to extend phonemes otherwise almost entirely restricted to Arabic loans into inherited Berber vocabulary.

If you want the full story, go read our article: The Origin of Mid Vowels in Siwi (published in Studies in African Linguistics 45:1-2 (2016), pp. 189-208).

No comments: