----- Original Message -----
From: "Miguel Carrasquer Vidal" <miguelc@...>
>>You called it -*yH2eH1, which amounts to -*a:.
>>Can you clarify this at all?
> The declension of the a:-stems is usually reconstructed more
> or less as follows:
> Nom *-a: (= *-eh2)
> Skt. -a:, Lith. -à/-ó-, Slav. -a, Goth. -a/-o:, Grk.
> -e:/-a:, Arm. -0, OIr. -0. Latin is irregular with short -a
> (as in the n.pl.).
You previously stated that Latin had the weak allophone of alternating
so it's not so irregular ?
> Acc *-a:m (= *-eh2-m)
> Skt. -a:m, Lith. -aN, Slav. -oN, Goth. -a/-o:, Grk.
> -e:n/-a:n, Lat. -am, Arm. -0.
>Old Irish is irregular with
> -YN (< *-en). Note that *-m is consonantal here and in the
> Apl. (we do not have *-eh2-m., *-eh2-n.s)
Western PIE does not accept pre-glottalized phonemes.
so H2m was impossible in Celtic (or Italic)
No wonder Irish has **eH2m > *em
H2 was erased in that position.
Regular again ?
> Voc *-a (= *-e[h2])
> Lith. -a, Slav. -o, Grk. -a. The laryngeal coloured the
> vowel before it was dropped. Sankrit -e (*-ai) is irregular
> here, more below.
Not so sure
Sanscrit (and other Central PIE languages) may have fused :
eH2 > a and ez > ay.
Might not be irregular.