🚀 go-pugleaf

In-Reply-To: <VEM090L@minnie.zdv.uni-mainz.de> knappen@obelix.kph.uni-mainz.de (Joerg Knappen)

> It is quite easy to explain that two formerly
> different sounds merge into one, but that one sound in the _same environment_
> splits cannot be explained.

I agree that in classical comparative linguistics this would be
inexplicable, and one would have to search for unsuspected
environment differences (cf. Verner's Law). Or grammatical function
might be important - though some might argue that that is
environment, too.

But we now know, thanks to the sociolinguists, that although sound
changes may well end up without exceptions, they may proceed not only
environment by environment, but even word by word. Look at the Ripuarian German diale
cts, which have
unshifted Gmc t in `dat' and `wat' but all other Gmc t > s,
including, I think I'm right in saying, `etwas'. One might debate
about the reasons for this particular example, but in principle
either partial change or influence of neighbouring dialects would
account for this offence against neo-grammarian sound laws.

Peter
----------------------------------------------------------------------------
Peter Christian                         
  Dept of European Languages              peter@gold.ac.uk
  Goldsmiths' College, London.            petex@cix.compulink.co.uk
----------------------------------------------------------------------------