On Fri, 06 Dec 2002 09:16:19 +0100, Miguel Carrasquer <
mcv@...>
wrote:
>On Fri, 06 Dec 2002 06:36:12 +0000, "Glen Gordon"
>>They aren't. There is no **eg because the universally accepted
>>reconstruction is *ego:.
>
>No it isn't.
>
>Beekes: *h1eg^, *h1egoH, *h1egHom
>Rasmussen: *eg^
>Szemerényi: *ego:, *eg(h)om
>Schmidt (in Pokorny IEW): *eg^hom (n.) < *e- + *-ghe + -om
>Cowgill: *eg^
>Adams/Mallory (EIEC): *h1ég^, emphatic *h1eg^óm
>Adrados: *eg
>Lehmann: *egh
Let me expand a bit more on this. As you can see, only Szemerényi
reconstructs (non-laryngeal) *ego:, although he does not try to pass
it off as the sole reconstruction for PIE, of course. That is because
*ego: (or *h1egoH) simply does not qualify as a valid PIE
reconstruction.
A reconstruction, and I can see by your acc. **menem, gen. **menese
that you're having some problems with the concept, is required to
provide a valid path to all of its descendant forms in the daughter
languages, and there is just no way to get from *ego: to Baltic *es^,
Slavic azU or Indo-Iranian *ajham. That's why Szemerényi is forced to
provide an alternative form *eg(h)om. And even then it's doubtful
whether both reconstructions together can account for Baltic *es^ or
Armenian es.
=======================
Miguel Carrasquer Vidal
mcv@...