On Thu, 29 Jan 2004, tgpedersen wrote:
>
> I don't know why I get this feeling that you are being sarcastic.
I was.
> Let
> me try to be more precise. Palatals can be depalatalised only if they
> occur in alternating paradigms somewhere in the language, in which
> case they _might_ spread to non-alternating environments. But without
> the alternation depalatalisation won't get started in the first place.
Now that makes the theory a bit easy, doesn't it? Any conditioned change
is liable to produce alternation, so how do you get a contrasting case
without alternation (*anywhere* in the language, as I now understand it
means) showing that lack of alternation blocks all atttempts at
depalatalisation?
Actually, I can easily imagine how speakers can change their pronunciation
under the influence of that of others. I believe this is a more potent
factor governing the course of language change than any amount of
paradigmatic alternation.
In the case of Danish it was certainly not levelling that caused
depalatalisation. The correspondence of German <gelten, galt> used to be
<gjælde, galdt>; that developed into <gjælde, gjaldt> by levelling, then
to present-day <gælde gjaldt> with depalatalisation *against* the uniform
picture caused by the earlier levelling.
Jens