> ====================
> H2 stands for 8 proto-phonemes.
> 4 are voiced :
> *& *z *dz *h
> 4 are glottalized.
> ? s? ts? th?
>
> Phonotaxis of proto-Indian
> not to be confused with indo-iranian
> has is that t/k/p followed by any H :
> yields th kh ph
> but this cannot be projected into PIE
> It applies to Proto-Indian alone.
Is there any internal evidence for this eightfold division of *H2?
============
Yes
A lot.
I suppose "internal" means PIE
without N Caucasic and without Yenissei.
1. &ayin
This is the one in Hittite <h>
It also leaves aspiration.
normally after consonants : C+ & > Ch
Sometimes after :
Greek laas "stone" < *l_&
Greek lithos < *li&t-
This is not the regular treatment in Greek (Substrate).
2. 3. z and dz
It's hard to distinguish the two.
The pattern is
Eastern PIE : z > y
Western PIE : H2
Anatolian : s ts
Very Often,
you find (fossil) words in Romance languages
that have s : ak- "stone" but Latin sak-sum
av-esna but also sabaia < *z_b
Salpuga : Arakhnê < z_r
Cf. FRench isard "mountain goat"
Cf. Semitic &_z "goat"
Each time you have s versus H2
Look for z or s? in external data.
Greek is highly erratic :
zab > Zeia
dza?-p "upupa epops" > H1ep !!!
Hesychius apupa is less strange.
zak? "sacred" > yag-nos (regular)
zahva "sea" > Aigian sea (borrowed)
An example of *dz is causative
Tokharian -s-
Eastern PIE -y-
Anatolian -zz-
This is not three morphemes
but three reflexes of one phoneme.
Orthodox PIE phonology is stupid
to the point of
not seeing what PIE morphology is.
4.5 s? ts?
Same pattern as with z dz.
Most often
the "Erweiterung" -s- after H2 is useless.
It's *s? dismembered into H2 + s
Stupid again and uselessly complicated.
Sibilants z dz s? ts? are often kept as s-C
because of phonotaxis.
I suspect Tokharian and Germanic have kept
s? and ts? as -s-
because vowel-s? evolved into v-?s > v?-s > v:-s
Like in the word to flare < bhleH2-s- < *bhle-s?-
A complete reassessment of H2-s- is necessary.
6 ? glottal stop
Causes irregular voicing.
Pattern
Celtic ?+C -CC- (unvoiced geminate)
Eastern and Central PIE ?-C > voiced
7. h
this one sometimes hardens into -k-
as in *zahab "gold"
Lituanian au-k-sis.
8. th? (interdental emphatic)
A rare and difficult phoneme
This one sometimes surfaces as -dh-
instead of H2
like in Gothic ed remade
from *b_y_th? "egg"
Ari-th-mos is also a case.
Arnaud
================