Re: [tied] Re: PIE genitive plural *-o:m, a possible analysis

----- Original Message -----

From: proto-language@...

To: cybalist@yahoogroups.com

Sent: Tuesday, April 25, 2006 11:06 AM

Subject: Re: [tied] Re: PIE genitive plural *-o:m, a possible analysis

 ***

Patrick:

 

Like Torsten, I am very skeptical of any explanation which is made by appealing to "analogy".

 

"Analogy" is basically the idea that a form which is not to be expected by applying acknowledged rules has been mistakenly patterned after a different form which is expected. While I certainly concede that this is "possible", I doubt whether it is as common as many would propose. Furthermore, it cannot really be proved.

 

I would like to try to explain some derivatives of *wékW- by means of the assumptions I have detailed in previous postings.

 

If Piotr or anyone else feels that these assumptions are wrong, I would welcome the criticism so long as it is explanatory rather than categorical.

 

A common PIE suffix indicating 'quality' or 'state' is -*sA, which creates neuter substantives.

 

Combined directly with*wékW-, the rule calling for penultimate stress-accent, and deletion of stress-unaccented vowels, would produce *wékW-sØ.

 

-*sA could also be combined with the plural form, which was formed by shifting the stress-accent one position to the right: if done during the earliest stage, this would produce *ukWé-sØ; if done during the next earliest stage, this would produce *wokWé-sØ. During this stage, *é became *o when the stress-accent was removed.

 

After the significance of the thematic extension was forgotten (plurality), the stress-accent was retracted to the root to resume the (singular) penultimate stress-accent: *wékWo-sØ.

 

At a yet later stage (Beekes' third stage), *o could be stress-accented and not revert to *é but rather became *ó; from *wokWé-sØ, this would first yield *wókWo-sØ.  The next operation was that the stress-unaccented *o was deleted, producing compensatorily *wó:kWØ-sØ; but defining a rule for this deletion is problematical. I would be glad for any constructive suggestions on the formulation of this deletion rule.

 

In PIE (though not Indo-Iranian), _all_ lengthened vowels are compensatory for the loss of aspiration, a laryngeal, or any other consonant or vowel.

 

In connection with these forms, we should remember that the earliest PIE "genitive" was 'relational', i.e. -*yA, an inheritance from PAA; it survives as a PIE locative though, of course, some PIE-derived languages retained it.

 

***