[tied] Re: On the ordering of some PIE rules

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:

>
> On 2007-06-07 21:33, tgpedersen wrote:
>
> > Did you want to comment on my proposal for PIE spirantization or
> > not?
>
> Not in the last posting, but in one of the earliest ones I admitted
> the possibility that *p and *k had fricative allophones (before *t
> and *s) well before Grimm's Law.

Throw in *t and a rule which turns spirants into sibilants (ð > z, þ >
s) before stop and you can give up complicated affricativizations of
dental + dental to explain West-IE -ss-, East-IE -st-.

A few details: the verbal adjective that does service as ppp is a
thematic end-stressed adjective: *<stem>-tó-. They come in pairs
(tómos/tomós) with stem-stressed nouns. According to me, both are
generalizations of a non-thematic nom. *<stém>-t-s, gen. *<stem>-tó-s.
I think that in those languages which had a phase of initial stress
(West-IE), the paradigm of stems ending in dental was generalized on
the nom., which was initial-stressed already, with the result
*<stem>-ss-o-, in the rest the gen. etc was generalized, with result
*<stem>-st-ó-.

> This could be a feature linking Germanic with Celtic, at least
> areally (but hardly with Iranian, which had spirantisation in a
> different context).

Iranianian spirantization like Germanic, Sabellic and Old Irish leaves
the second of two stops untouched. Other that that, it is more
extensive that that of Sabellic and Celtic, but less so than that of
Germanic.

> I doubt , however, if anything of the sort can be proposed for PIE.
> You'd have to propose massive despirantisations not only for Baltic,
> Indic or Latin but also e.g. for Greek and Hittite.

The result is despirantization, but the mechanism is regularization.

> That's more expensive than "no change at all".

Despirantization is. Regularization isn't. BTW I hear that some
Russians are starting to collapse the g/z^, k/c^ alternation in verbs?

> Both pre-obstruent lenitions like pt, kt > ft, xt and fricative
> dissimilations like xT, fT > xt, ft are fairly common changes
> cross-linguistically. Both of them have taken place in colloquial
> Modern Greek, for example:
>
> ptero- > ftero 'feather'
> hepta > efta 'seven'
> okto: > oxto 'eight'
> kHtHes > xtes 'yesterday'
>

I don't doubt it, but in 4 separate branches?


Torsten