Re: [tied] The disappearance of *-s -- The saga continues

> >> I started as retroflex /s./ but now I would say it is generally
> >> /S/, at least in Central Sweden.
> >
> >Ugh, more complexity. Well, if this helps, my greatgrandparents
> >came from Stockholm around 1900-something. Something tells me
> >that this [s.], even if it is retroflex isn't terribly so.
> >Otherwise, I'd imagine it sounding more Hindi :) All I remember
> >is a "softening" of /s/ to what sounded like an "s" with a
> >tip or flat tongue further behind the teeth than normal but only
> >by a few millimeters or so. I suppose that can be called
> >"retroflex" but not very much and the tongue wasn't curled back,
> >I don't think.
>
> It was originally as "retroflex" as Polish sz or Mandarin
> sh, i.e. not really retroflex at all (Ladefoged uses the
> term "laminal flat postalveolar"). The acoustic effect is
> similar to that of retroflex /s./, and the conditioning in
> Swedish by /r/ (as with the stops: /rt/ > [t.]) is another
> common factor.
>
> The retroflex shibilant /rs/ pushed old Swedish /S/ (<sj>)
> into the sound it has acquired now (a labiovelarized
> labiodental or dorsovelar fricative), which means that /rs/
> can now move into the territory of phonetic [S].

As a native speaker I can confirm all of this. Swedish /rs/ is like
polish /sz/ (not like Hindi /s./) and is conitioned by an r in the
preceding consonant cluster (even happens across word bounadries).
It doesn't happen at all in some dialects like east-Swedish, i.e the
Swedish spoken by Finns, who have -rs- as trilled r plus s (and -rd-, -rt-
preserved as well). What happens with the old Swedish /S/ (<sj>) is like
Miguel says, except in the north (say north of Sundsvall) where it remains
/S/.

mvh

Harald