Re: [tied] The Precise and scientific nature of PIE reconstruction.

From: mkelkar2003
Message: 36993
Date: 2005-04-08

--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
> On Thu, 07 Apr 2005 21:15:45 +0000, mkelkar2003
> <smykelkar@...> wrote:
>
> >Here is how PIE reconstruction works as explained by McWhorter
(2001,
> >p. 45-46). He gives seven IE cognates for the word "sister-in-
law".
> >
> >Sanskrit: snusha
>
> snus.á:
>
> >Greek: nuós
> >
> >Old English: snoru Armenian: nu
> >
> >Russian: snokha Albenian: nuse
> >
> >Latin: nurus
> >
> >In Armenian and Albenian the word actually means bride not
> >sister-in-law. He explains this phenomenon as an example of
semantic
> >change.
> >
> >The PIE word is reconstructed as follows:
> >
> >1. The word should begin with sn rather than n. The missing s in
the
> >languages has been lost due to attrition.
>
> *s- is always lost in Greek and Armenian. In Albanian and
> Latin sn- > n- is regular.
>
> >2. The first vowel must be u rather than an o. Russian and Old
> >English have muted that to an o. The majority rule applies here.
>
> No. It has nothing to do with majority. It has to do with
> knowledge about the languages in question. PIE *u gives
> Slavic /U/, and that's exactly what we find in OCS snUxa.
> The hard yer regularly gives zero or /o/ in Russian.
> In Germanic, PIE *u gives /o/ or /u/ (PIE *o gives /a/),
> depending on the front or backness of the following vowel.
> *snusa: regularly gives *snora:.
>
> > So far the proto word is *snu
> >
> >3. The next consonant should be an s as shown in Sanskrit,
Albenian
> >and Greek. S has mutated to a k in Russian and r in Latin and Old
> >English. Therefore, *snus
>
> Actually, /s/ gives /s^/ after *u in all the satem
> languages, so we have snus^a: in Sanskrit *snus^a: > snUxa
> in Slavic.
>
> Intervoalic /s/ gives /z/ and then /r/ independently in
> Latin and Germanic.
>
> >4. So far so good. The most twisted part is the ending. The
majority
> >of the above seven words have a masculine ending except for
Sansrkit
> >and Russian.
>
> No. Old English has a feminine ending too, and so does
> Albanian.
>
> >So the original PIE word MUST be masculine! and end with
> >an os rather than an "a" a feminine ending, as it does CORRECTLY
in
> >Sanskrit and Russian. Therefore PIE *snusos. Science does not work
> >by the rule of the majority.
>
> No, science works by logic. Disregarding the Albanian word,
> the ending of which is not secure, we have the following
> situation:
>
> Sanskrit, Germanic and Slavic point to *snusá: (f.)
> Greek and Armenian point to *snusós (m.)
> Latin points to *snusú:s (f.)
>
> The Latin ending is analogical after socrus "mother-in-law"
> (*swekru:), and the pre-Latin form was doubtlessly *snusós,
> as in Greek and Armenian.
>
> We can now choose to reconstruct *snusá: or *snusós. If the
> proto-form was *snusá:, we have to explain why this word,
> denoting a female person, acquired the normally masculine


The NORMALLY masculine ending! Gotcha! Obviously what is "normal" is
decided by majority rule, which is fine with me as long as linguist
admit that, and not prentend to do science.

Let us mess up this pretty picture of language families. There might
be another word in say Swahili that must be included in the
reconstruction also. You must check every single living or dead
language ever spoken on earth and that has a cognate include that in
your families.

Greek, Albanian, Armenian and Latin could be a "family" because they
don't have the beginning s in the word. Sanskrit and Russian can be
a family for they have the ending correct.

WHERE these languages are spoken TODAY and by WHO should play no
role in how the language families are constructed.

M. Kelkar



> ending *-os in Greek, Armenian and Latin. This goes against
> all logic. If the proto-form was *snusós, on the other
> hand, it is trivial to explain why the ending *-os was
> replaced by feminine *-a: in Sanskrit, Germanic and Slavic.
>
> So logic dictates that the original form was *snusós.
>
>
> =======================
> Miguel Carrasquer Vidal
> mcv@...