Re: [tied] Re: *kW- "?"

> --- In cybalist@yahoogroups.com, mkapovic@... wrote:
>> > So the anomalous languages are Albanian, Luwian and Aremenian, all on
>> > the southern fringe of the satem area, in which the merger *kW > *k
>> > did not occur immediately.
>>
>> Sorry, Richard, but I seem to have missed the point... :-/ Could you
>> elaborate?
>
> Geography:
>
> I don't think Albanian and Armenian need any elaboration. For
> Anatolia, I see the geographical connection between Anatolian and
> Indo-European proper as being across the Bosporus. The precise
> location of the interface is, I think unclear. However, it is the
> West of Anatolia that connects with south of the area of IE proper.
>
> It is possible that there may not have been a connection - Phrygian
> and possibly Tyrrhenian complicate the picture. Phrygian is a
> langauge of uncertain satemisation - perhaps it got only as far as
> Stage 1 below.
>
> Dialect Geography:
>
> As I understand it, sound changes are not necessarily restricted to
> languages, but may spread across to neighbouring languages. A large
> scale example is the loss of voicing contrasts which has swept through
> much of China and mainland SE Asia (typical giving tone splits, but
> sometimes merely a register distinction), though there are various
> hold-outs - some Yi dialects have resisted it, and it is only now
> affecting the Li language of Hainan.
>
> Satem Development:
>
> Three basic stages:
>
> 1. *k^ [k] palatalises [c]?, [k^].
>
> 2. *kW > k, merging with PIE *k. However, other changes may bleed
> this development, leading to the divergent development in Albanian and
> Armenian. There is nothing striking about this development if
> previous [k] has been depeleted - consider the developments of Romance
> [kw] to /k/ in French and Spanish. The change may not need such a
> motivation - such a simplification in Farsi as opposed to Daric
> Perisan may not have had such a motivation.
>
> 3. [c] or [k^] affricates.
>
> I am not sure that we actually know the relative order of stages 2 and 3.
>
> Thesis:
>
> Stage 1 occurred over a wider area than Stage 2. In particular, Stage
> 2 did not extend to the far south of the are affected by Stage 1.
> Stage 2 may have occurred independently significantly later - after
> other changes had bled it.
>
> Thought: Was Stage 2 (generally merger of *kW and *k) actually a
> connected phenomenon? Parallel developments seems quite common.

OK, but how can you still derive all of this from only 2 original velar
series? If I'm not mistaken, that was your basis.
The fact is that Luwian shows different outcomes of *k, *k' and *kW which
means that it was the same in Proto-Anatolian and in Proto-IE. It's just
that Luwian (and Lycian) are the only IE lgs that preserve the distinction
in all environments (but only in voiceless series), Arm.&Alb. preserve it
only in front of front vowels and all others preserve just two series (the
traditional centum-satem lgs). You may claim that Luwian etc. are residues
which preserve the older stage of all satem-lgs but that doesn't change
the fact that we have to reconstruct 3 velar series for IE, in spite of
the fact that many scholars don't like that.

Mate