Re: [tied] Stop horsing around

--- In cybalist@..., "Glen Gordon" <glengordon01@...> wrote:

From: "Glen Gordon" <glengordon01@...>
Date: Tue Apr 30, 2002 8:32 pm
Subject: Re: [tied] Stop horsing around

> Elaborate. What possible reanalysis are you speaking of? I suppose
if I were an AA speaker listening to a speaker of a Satem dialect say
*ec^was, I'd probably hear it as **a-c^wase with a demonstrative
marker *a- attached.

That's what I mean. The AbAd languages (I prefer this abbreviation,
because Afroasiatic is lurking round the corner) -- so, the AbAd
languages love prefixes. In Modern Abkhaz 'the horse' is <a-
c^:&>, 'my horse' is <sarra s-c^:&> 'your(f.) horse' is <bara b-c^:&>.

> But still, the nominative or accusative, even the locative, would
be most usually heard with the stem *ec^wa- and you'd think that this
would be fused to the AA loan. The vowels don't even match either.
How does *[a] become *[&] when AA speakers could surely
have told the difference?

I know too little about early AbAd to speak with confidence about
patterns of substitution in protolanguage loans, but these particular
substitutions are not unprecedented. Indo-Iranian short *a was not
necessarily fully open, judging from its historical reflexes. At any
rate, we have Circassian <s'&> and Kabardian <s^&> 'horse'
corresponding to a later stage of Proto-Iranian (*as'wa-, with a
fricative). Finally we have Khinalugh <spa> 'colt', this time sans
doubt from still later (dialectal) Iranian *aspa-, and for all I know
the Dagestanian (and some other) forms may be even more recent (cf.
Khinalugh ps^i : Ossetic aefs^ae 'mare' from metathesised *aspa:).
The correspondence problem gets even worse, anyway, if you want to
make *ek^wos a loan from Proto-AbAdHa. How do you derive it in _that_
direction? As far as I can see, there's no basis for reconstructing
a 'horse' word for the whole AbAdHa group in the normal comparative
way (I'm not even sure if a common AbAd reconstruction is fully
justified). To begin with, you haven't shown any evidence that the
Proto-AbAdHa sound corresponding to AbAd *c: (or whatever the exact
affricate in the 'horse' word) goes back to an older (palato)velar --
it's mere speculation.

> Finnish and I had a chat and she tells me that the *-s isn't
as "weak" as you suggest:

> porsas "pig"
> taivas "heaven"
> marras "dead"

A very good point, Glen. I should have thought of that -- silly me. I
must admit that my suggestion doesn't work for archaic Indo-Iranian.

> Finnish also has many examples of a fused IE nominative in other
loans:

> hammas "tooth"
> kuningas "king"
> ruhtinas "prince"

> So if every other language borrows words complete with inflection,
why is Abkhaz-Adyghe so special??

Have you really examined "every other" language? "Every other" in the
sense 'all the other' or '50% of' languages? It isn't difficult to
find examples of regular inflections getting systematically deleted
in loans. Even English often drops Latin -us, e.g. sine < Neo-Lat.
sinus, and that despite Fr. sinus; cf. also Pliny, Antony, Livy, etc.
with -y for -ius.

Piotr