----- Original Message -----From: Richard WordinghamSent: Tuesday, October 24, 2006 6:35 PMSubject: Re: [tied] Origins of Indo-European, and naturalness of laryngeals--- In cybalist@... s.com, "Patrick Ryan" <proto-language@ ...>
wrote:
> While I do not doubt that any consonant in any language has some
small allophonic effect on any vowel with which it comes into contact,
as Brian concisely stated, the question here is whether an allophone
of an existing phoneme can achieve new phonemic status; and maintain
that status even after the elimination of the conditioning consonant.
I trust you don't mean what you are saying - that would cast doubt on
the basis of Germanic umlaut.***
Well, that is certainly a dragonfly in the ointment.
I guess I should have said "direct contact".
Umlaut seems to me, as I understand it, to be "indirect contact": CVCi/y, for example where i/y is the conditioning factor (as opposed to CVC where the final C is the conditioning factor). It is anticipatory assimilation.
Also, I wonder about the phonemic status of, say, German <ö>. I do not believe there is a native word in MHG in which CVC contrasts with CöC when CöC does not derive from an earlier CoCi/y. Is it possible to say that <ö> is simply an indicated allophone of <o>?
***
I hope you are merely saying that you don't believe a consonant could
condition a split of a vowel phoneme into two vowels and then vanish,
not that there is convincing evidence that laryngeals created new
vowel qualities - /a/ and /o/ seem to have origins besides laryngeal
colouring of /e/.
The best example I can think of is the split of French /a/ into a back
vowel /A/ before /s/ (e.g. _pas_ /pA/, _pâte_ /pAt/) and a front vowel
otherwise (e.g. _chat_ /Sa/, _patte_ /pat/).***
I do not know enough about French to really dispute this with any certainty.
But I would be interested to know if French /A/ existed in circumstances where being before a former /s/ was _not_ a conditioning factor.
Patrick
***