From: Piotr Gasiorowski
Message: 13435
Date: 2002-04-23
----- Original Message -----From: tgpedersenSent: Tuesday, April 23, 2002 1:07 PMSubject: [tied] Re: Satem> I still think it phonetically odd that the satem languages, in contrast to the Romance ones, palatalize also before back vowels and even consonants. Is there an explanation for that? Is there even a parallel in the world?There are some parallel examples, e.g. the case of Salishan in North America: in fourteen Coast Salishan languages and two Interior Salishan ones etymological non-labialised velars (preserved in the rest of Interior Salishan) changed into alveopalatals in _all_ environments. Interestingly, in Salishan the velar gap was _not_ filled by labiovelars or uvulars (both available in those languages), so that in some of the Nortwestern Coast languages plain velars occur only in loanwords (a possible explanation in terms of systemic pressure is maximal differentiation in a particularly rich inventory of dorsal consonants). The case of Satem IE would have been less unusual because the delabialisation of *kW restored the "naturalness" of the system. Here's an interesting comment by Sarah Thomason:"... More interestingly, when people talk about unconditioned sound changes they are referring to the end point only: there is no claim, implicit or explicit, that the change began everywhere and proceeded through the lexicon randomly. So, for instance, to say that most Salishan languages (Pacific Northwest, U.S. and Canada) underwent a change from nonlabialized velars to alveopalatals does not imply a claim that the change happened simultaneously in all environments. It seems most likely, in fact, that it began, like other palatalization changes, before /i/ only, or before front vowels. (I should have said: like most other palatalization changes; not all.) And then it generalized until all the nonlabialized velars were swept away, so that Flathead (e.g.) has no plain velars at all, except in two recent loanwords ("coffee", "coat"). But since the initial stages of this change are not documented in any way, there's no way to test the hypothesis that the change began as a conditioned change."
Piotr