Dear Sabine, kalê hêmera,
> I can't follow this - may be I just don't understand what they were
talking
> about? If PIE had the glottal consonants Gramkrelidse/Ivanov suggest,
> wouldn't it follow that it didn't have voiced consonants? For my
logic that
> would make Lin. A a candidate possibly closer to PIE (which - taking
into
> account how old it is - would make sense), wouldn't it?
>
In fact, in G&I's theory, there are still voiced consonnants, beacuse
there were three series of stop : the voiced aspirates of the
traditional theory are rteconstructed as plain voiced stops (or, to
give justice to their theory, they postulate two allophones, aspirated
and non aspirated, for both voiced and unvoiced stops). So the
traditionnal bh dh gh p t k are creconstructed b(h) d(h) g(h) p(h) t(h)
k(h).)
I think that the absence of voiced stops in a language does not
disprove it is IE : icelandic has no real voiced stop, and according to
Ladefoged & Maddieson 1995, neither does english. Anyway, the
voiced-unvoiced axis is not discrete, (at least from a phonetician's
point of view) so this opposition can be realised in an infinity of
different ways according to languages.
I think the point with Lin A is that is has (if I remember well) only
ONE series of stops : that is exceedingly rare in IE languages. I don't
know much about hittite, but, as far as I am aware, there was only one
series of stops also. Many words can be spelled with both voiced and
unvoiced obstruents in the akkadian syllabary. However, it might just
be because the voicing in anatolian was so different from akkadian that
hittite scribes couldn't prceive the voice opposition in akkadian.
Anyway, I am afraid the problem of Lin. A.'s genetic affiliation is
insoluble. It is exciting to think it was an isolated language.
Guillaume