Re: Sos-

A pre-(not Proto-)Gmc double root *skran,W-/*skran,gW- (not too
unlikely) would (might?) produce Gmc skrov-/skronk-.
But to get the Bornholmsk, French etc, we'd need something like
*(s)kort(s)-/*(s)korl-.
I can't get the thought out of my head that it might be related
somehow to the Uralic *sose- (*sase-) root by cluster reduction, which
would be what Uralic does.

Is there any particular reasion why the Uralic s´, I think it is,
which so weirdly turns into s-, t-, l- wasn't a cluster instead, say
*skr-/*str-?

Torsten

======

The exact nature of the opposition between proto-Uralic *s and *s' is not
clear to me.
The first point is it's standard PU *s that has this s / t / l / y
alternation,
whereas *s' is usually stable as /s/ or /s'/ or the like.
Maybe this ultimately reflects a contrast between *z and *s > PU *s and *s?
(glottalized) > PU *s'.
but this is only the beginning of a guess...
I don't have a good deal of clean examples.

A.