On Fri, 07 Feb 2003 13:32:37 +0000, "Glen Gordon"
<
glengordon01@...> wrote:
>Miguel:
>>Well, it's not as simple as that, far from it. The problem of /a/ is
>>a tough and subtle one, and simple rules do not apply.
>
>Occam's Razor. Unless you can support the need for a more complex
>theory (which you have not yet done), a simpler one should suffice.
>
>You initially claimed that nasalization is the cause of *a in a select
>list of morphemes where laryngeals appear to not be the cause. Both *m
>_and_ *n are nasal, so they must be both included in your nasalization
>rule.
>
>You now claim that you're "pretty sure it doesn't apply to **na:",
>showing that nasalization cannot be the cause, or sole cause, since *n
>is quite obviously a nasal phoneme.
But /m/ and /n/ are not the same phoneme, and from the point of view
of historical linguistics, I would claim that /m/ is "more nasal" than
/n/ (I'm not sure how that translates into phonetical terms).
Historically, /n/ is often denasalized (yielding /r/, /l/ or /d/ in
the history of e.g. Albanian/Romance -n- > -r-, Proto-Indo-European -n
> -r, Sumerian n ~ l , etc.) Such developments are much less common
in the history of /m/.
On the other hand, final -m after a vowel is often lost (presumably
passing through a stage with nasalization of the vowel), whereas -n is
more stable (cf. Latin where -m was lost everywhere (except in
monosyllables, where it became -n), but -n largely retained (except in
Romanian)).
>This conclusively shows that your initial "nasalization claim" is either
>incomplete or incorrect.
It's incomplete, as I have always said. Occam's razor was never meant
to force us into accepting simple but incorrect theories over complex
but correct ones. The attempts at explaining PIE /a/ through
laryngeals only have failed. A good number of the non-laryngeal cases
can be explained by the influence of uvular *q, *G, *Gh = *k, *g, *gh
(kap-, ghabh-, kar-, etc.). That still leaves a number of unexplained
cases. My nasal theory can take care of most of them, if only I could
find the correct (i.e. complete) formulation of the soundlaw(s)
involved. That nasalization applied widely (also to short and
unstressed vowels) in some stage of pre-PIE is demonstrated by the
exceptions to the -n > -r rule (*-men does not yield *-mer). The
problem is finding under which exact conditions nasalization was
completely lost (V~ > V) [V is short/unstressed?], under which
conditions it was reanalyzed as a consonantal feature (V~ > Vn/Vm)
[before /s/?] and under which conditions it affected the quality of
the vowel (V~ > V') [V=/á:/, V'=/á/ after /m/, before /nC/ ?].
=======================
Miguel Carrasquer Vidal
mcv@...