--- In
cybalist@yahoogroups.com, Piotr Gasiorowski
<piotr.gasiorowski@...> wrote:
>
> ----- Original Message -----
> From: tgpedersen
> To: cybalist@yahoogroups.com
> Sent: Thursday, May 22, 2003 10:54 AM
> Subject: Re: [tied] kentum and satem
>
>
> > I know you believe there is some ontological difference between
> garden variety palatalisation plus paradigm regularisation on the
one
> hand and satemisation on the other. I know we discussed this
before,
> but what are your reasons for assuming that? Or, stated
differently,
> if the Phrygians had regularised their paradigms (in the satem
> direction) in what way would Phrygian then not have been a satem
> language?
> The difference is simple -- that between a conditioned change (like
Verner's Law) and a context-free one (like Grimm's Law). The Satem
shift shows no trace -- let me repeat -- NO TRACE of palatal
conditioning of the "garden" variety.
What kind of trace would you have mind here?
The *k^ series sounds were fronted in ALL positions, also before back
vowels and consonants, where no "paradigm regularisation" had a
chance to occur. The CONDITIONED palatalisation of *k and *kW stops
also took place in several Satem branches, but its effects were
different and despite paradigmatic levelling-out (as in Sanskrit,
where there was a lot thereof) there is enough evidence of the
original conditioning context. A Satem language is one in which *k^
is fronted not only before front vowels (or before apophonic *o
alternating with *e, where analogy might have been at work), but also
in words like *h3ok^to:, *h1ek^wos, *k^lewos, etc., where analogy can
be ruled out.
>
As you may have noted in my 'shibbolethisation' sketch, there are
examples of regularisation caused by a shibboleth pair also outside
of paradigms? (Sw. 'djur', Da 'dyr' "animal")
Torsten