From: "Pavel Iosad" <[log in to unmask]>

| And, off topic rant, has anyone heard of a second reinterpretation, designed
| to undermine the drawback of the GT itself. I have a book by Starostin who
| says there's a possibility of the three series being unvoiced, unvoiced
| geminates and voiced. This is also typological, and has the virtue of the
| possibility of voicing of initial geminates (suchn things happen, while
| voicing of glottalized initials does not).

[Of course GT means Glottalic Theory (of Proto-Indo-European).]

Paul J. Hopper interprets traditional (that is, Lehmann's) T, D and Dh (T =
voiceless stop, D = voiced stop) as T, T` (glottalized) and D_ (D underlined,
meaning "murmured"), a compromise view of Gamqrelidze-Ivanov that preserves the
aspiration of the traditional voiced aspirated stops. (Hopper is just as much
responsible and deserving of credit for GT as Gamq.-Ivanov.)

An older proposal was by Joseph Edmonds, where T > Th, D > T and Dh remained Dh.
The plain voiced stops are devoiced and the already-voiceless stops are

Holger Pedersen reversed the voiceless and voiced: T > D, D > T. I don't know
what he said about voiced aspirates. His logic was based on the rarity of
traditional *b, and it would make sense that *p would be missing instead of *b
in a natural language, which common Indo-European indeed was aeons ago.

Oswald Szemerényi believed in four classes of stops instead of three: T Th D Dh.
His theory today is largely rejected because of the rarity of voiceless
aspirates and the over-reliance on Sanskrit data.

I could take a position that T > Th, D > T and Dh > D (with an allophone of Dh),
where the (new) plain voiceless stop need not be ejective or otherwise
glottalized, but may be "half-voiced". Another theory of mine is in a parallel
to Korean and Middle Chinese, where the triad is, in order: voiceless aspirated,
plain "lax" voiceless with voiced intervocalic allophone, and the "glottalized"
consonant is now the voiced (aspirate) stop, which parallels the "tense"
"ssang-" Hangul consonants. So then the result is T > Th, D > T/D, and D > DD or

The obvious point is that the PIE stops are much easier to describe phonemically
and typologically than phonetically.