Torsten,
They are lookalikes, not cognates. Cognacy
is demonstrated by establishing regular correspondences and showing that
your forms conform to them. Even if you want to get around this requirement by
postulating loans rather than cognates, you have to make it clear which ancient
stage of the source language the borrowing comes from. Your use of
ahistorical collections of modern words (following Manansala, in the case of
Austronesian), with no diachronic analysis, produces merely a mirage of
similarity. For example, Malayo-Polynesian "fire" words like <afi> or
<ahi> may seem to resemble PIE *h2ah-, until you realise that the medial
consonant reflects earlier (lenited) *p (Malay api might have given you a
hint). Actually, the PAN protoform is *Sapuy (> *hapuy > api etc.). The
similarity is gone. You don't go through this kind of historical checking branch
after branch. The fact that, say, modern Binandere has a "fire" word which looks
exactly like Sumerian izi (used almost 5000 years ago), means nothing. Sumerian
didn't borrow words from modern Binandere, and the early MP prototype of the
Binandere "fire" word was not like <izi> at all.
Piotr
----- Original Message -----
Sent: Wednesday, October 03, 2001 10:56 AM
Subject: Re: [tied] a(i)s-
Which would wrap up the case if *H2ay- had been an
IE-only word. But
it has cognates in AfroAsiatic and FinnoUgric (says
Bomhard) and
Sumerian and Austronesian (says Manansala), see
http://www.angelfire.com/rant/tgpedersen/Hs.html