Re: Kentum/satem (was: The future of Sturnan)

> Date:         Mon, 29 Apr 2002 10:24:16 -0400
> From: Muke Tever <alrivera@...>
>
> >===== Original Message From Constructed Languages List
> >- the satem languages constitute the more innovative group, that is
> >located more or less in the centre of IE territory;
> >- the kentum languages, located at the periphery, broke off earlier
> >from the family; they didn't take part in the innovations of the
> >satem languages and more or less preserved the original IE
> >phonology.
>
> As far as I understand it...
>
> In the centum languages, the PIE palatals (k' g' g'h) merged with the PIE
> velars (k g gh).
>
> In the satem languages, the PIE labiovelars (kw gw gwh) were the ones that
> merged with the velars.
This is if you believe in PIE plain velars. The jury has been out on
this matter for more than a hundred years, but no verdict yet.

If you don't, the statement is rather that the palatals became plain
velars in the centum group, the labiovelars became plain velars in the
satem group --- and in some positions, like after s-, both series
became plain velars in both groups.

> I don't think either is more or less 'preserved' or 'innovative'?
> Unless I have misplaced some point.
The labels do not refer to the phonological facts, but to the
distribution. It's commonly accepted that innovations usually spread
from a geographical point, so if you find a connected area with one
feature, and a disjointed fringe around it with another, the connected
area is assumed to be the innovative one.

In this case, a naive reading would have the innovation consist in
rolling back the palatal > plain change and doing labiovelar > plain
instead. Not very plausible.

However, I'm sure it's possible to set up more likely developments.
For instance, if there were only two original series, they could be
plain and labiovelar, preserved in the centum langs. The innovation
would then be a chain shift to palatal and plain in the satem langs.

Remember that the labels palatal and labiovelar for the two main PIE
series are somewhat arbitrary --- more or less a consequence of the
theory that there are three series.

Lars Mathiesen (U of Copenhagen CS Dep) <thorinn@...> (Humour NOT marked)

Reply