--- In cybalist@..., guto rhys <gutorhys@...> wrote:
> How would one explain that Breton has one (only one as far as I am
aware) exception to the rule st->s- which is 'sterenn' (star) cf.
Welsh 'seren'. Neither does Cornish, a more closely related form,
retain 'st-'. Could it be Latin influence or once again a borrowing
from a conservative dialect? Any suggestions as to why this one word
would behave irregularly.
There are a couple of other exceptions that I dug up by trawling
Pokorny (thanks for the link, Sergei!).
From IE ste/oub(h)ma:
Welsh 'ystum', Breton 'stumm' (bend)
From IE stomen:
Welsh 'safn', Breton 'staon' (palate)
Welsh 'sefnig', Old Cornish 'stefenic'
There were also a couple of possible Latin loanwords that did not
display st- > s-, e.g. Welsh 'ystawd', possibly from Latin 'status'.
The dialect mixture trick might work, but is less plausible. One
would be suggesting that a dialect with s-, sp- and sk- but no st-
would then borrow dialect forms in st- without simplifying them to s-
. (This scenario would be more plausible if Breton and Cornish
developed and then lost the epenthetic vowel, as Italian did. I am
not aware of any evidence that they did.)
An alternative is that old s- (now /S/, /h/ or /x/) had become so
distinct from the /s/ in st- that st- sporadically simplified to s-
with no pressure to restore the missing [t], and that this sporadic
change happened often enough to now (21st century) be seen as the
regular development. Possibly the exceptional words were exalted
enough (sterenn) - an effect Torsten has just mentioned - or rare
enough (staon) that they resisted simplification. Does this make
sense for 'sterenn'? I am no expert on Celtic culture.
I'll ask about this sort of development on PhoNet -