--- In
cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@...> wrote:
>> One has to respect the evidence in the first place, but there is
none of it for the conjectured ejectives.
> Substitute "occurrence in Armenian" and "Winter's law" and you have
the situation for glottalic stops. BTW the latter links glottalicness
and laryngeals, which links also their likelihood of diappearing from IE.
As for Armenian, or _Eastern Armenian_, to be precise, it has no true
ejectives but voiceless stops that may be weakly glottalised by _some_
speakers (see the detailed description in Ladefoged & Maddieson, who
basically treat these sounds as plain voiceless stops). There is no
proof that this pronunciation is old: it does not occur in Western
Armenian (actually, to embarrass the glottalists, some Western accents
have similar sounds as reflexes of Old Armenian voiced stops!). Like
/t/-glottaling in English, it may be a recent phenomenon.
As for Winter's Law, the fact that vowels, diphthongs and vowel +
sonorant combinations developed similarly before laryngeals and before
modal-voiced consonants is interesting but doesn't prove that the *d
series was glottalised.
To begin with, most of the "laryngeals" had nothing laryngeal about
them, and even when lenited they disappeared via an [h]-like glide or
a schwa-like vowel rather than a glottal stop. If e.g. preconsonantal
*er&2 > *e:r in Proto-Balto-Slavic, it was some kind of compensatory
lengthening (preservation of phonological "overweight") rather than a
glottal effect.
Lengthening before the *d series may have been of the same kind that
the phonetic lengthening of English "short" vowels before voiced
sounds (<bad> vs. <bat>). It took place before fully voiced stops as
opposed to breathy-voiced and voiceless ones before the merger of the
two series. Some English-speakers make no distinction between
intevocalic /d/ and /t/, tapping both, but they retain the
vowel-quantity contrast in <madder> vs. <matter>.
> > Even if the independent loss of the laryngeals in several branches
> was unusual (and I'd claim that the loss of dorsal fricatives of
> glottal glides is far more likely than the change of [t'] > [d]), we
> have independent evidence that it _must_ have occurred.
> >
> You would, wouldn't you? ;-)
Yes, I would. I've seen it happen more often, haven't _you_?
>
> Torsten