Re: Limitations of the compartive method

From: fournet.arnaud
Message: 52044
Date: 2008-01-29

----- Original Message -----
From: "Mate Kapović" <mkapovic@...>
To: <cybalist@yahoogroups.com>
Sent: Monday, January 28, 2008 10:59 AM
Subject: Re: [tied] Re: Limitations of the compartive method


>I guess, if the numbers are correct, that this is due to a high number of
> "dialect words", i.e. "North European" words which appear for instance
> only in Germ. and some of the other families like Celt., Ital. or BSl. but
> not elsewhere. There are many roots like that in Pokorny.
>
> Mate
=================
tsalam? t?ob

Yes,
One of the major flaws of orthodox PIE is that
it describes the post-PIE proto-Central dialect
that ultimately led to Greek, Balto-slavic, Indo-Iranian
in the first place.
It also sheds light on Germanic because Germanic
shares areal features with Central PIE.
This "central" bias is a problem when one wants to
compare PIE with something else (Semitic for example).

Latin with its perfective/imperfective two-way system does not
fit into the three-way system of Central orthodox PIE,
And Anatolian is worse.

Let's take an example : words for gold
PIE word : H2aHw-s
kept in LAtin aurum
CEntral PIE innovation : ghel-to
from the root *ghel "yellow".

If the word aurum were isolated,
it would have been discarded.
But the fact is this word exists in Hebrew :
zahab "gold"
Structurally zahab is H2_H2_w
I make you aware that Lituanian au-k-sis
with its strange -k- keeps trace of -h- in zahab
zahab > *HaHw > metathesis HawH-si > aw-K-si.

Pokorny is hugely biased and LIV is modernized
but it is not "improved" when it comes to describing
"real" PIE, not proto-Central-IE

Arnaud