Of cows and living

From: Piotr Gasiorowski
Message: 43340
Date: 2006-02-10

I've been thinking about the odd form of the verb 'to live' in PIE,
*gWíh3w-e/o-. It is usually thought to have been influenced by (or based
on) the adjective *gWih3wó-, but the relationship is strangely irregular
(why not a regular stative *gWih3we-h1-je/o-?), and if it is some
archaic (and otherwise unknown) kind of present stem derived directly
from the adjective without any suffixes, I still don't understand the
accent shift in the verb. I wouldn't expect any such thing, least of all
when the shift is to a nil-grade root syllable (cf. the <tudáti> type).

The accent shift, however, would be understandable in a reduplicated
thematic present of the type represented by *sí-sd-e/o- or
*stí-sth2-e/o-, where we have accent retraction in originally
trisyllabic stems (from something like **se-sed-é-). If so, *gWíh3we/o-
would make sense as *gWí-h3w-e/o-, with *h3 < *g(W) by simple
dissimilation. The adjective would then be deverbal rather than the
other way round, and its accent on the thematic vowel is normal. At the
moment I prefer to leave open the question whether PGmc. *kwikwa- is an
archaism or an accidental "evolutionary reversal". Suffice it to say
that the obscuration of the structure of *gWig(W)w-e/o- > *gWih3w-e/o-
resulted its reanalysis and the backformation of the "neo-root"
*gWih3(w)-, with its own retinue of derivatives.

If the original root was of the form *gWeu-, then in turn the 'cow, head
of cattle' word, *gWo:u-s, can be analysed as a corresponding root noun
(for the 'livestock' semantics, cf. *bHuh2-tlo- > *bydlo 'existence,
residence' --> 'cattle' in Slavic).