--- In
cybalist@yahoogroups.com, Andrew Jarrette <anjarrette@...> wrote:
> -- this is what Sanskrit is supposed to have had and what modern
Hindi, Panjabi, Gujarati, etc. have (though with velar
co-articulation), even though the many times I have heard the Indian
labio-dental approximant, and I have listened carefully and intently
with the intention of testing what I have read in books, I would still
swear that they use two allophones, a rounded one before rounded
vowels and perhaps back vowels, and a labiodental one before other
vowels, which are indistinguishable from /w/ in the former case and
/v/ in the latter case, to my ear at least.
That is entirely plausible. Thai is supposed to have the same
variation, but there's so much variation it's a statistical difference
rather than a pedictable allophone difference overall.
Richard.