Glen asked about Hittite pir 'house' and
its etymology.
The word is safely reconstructible to
Proto-Anatolian (with a number of derivatives in Hittite, Cuneiform and
Hieroglyphic Luwian, Lydian and Lycian). The most complete set of forms
occurs in Hittite; the other Anatolian languages contribute only auxiliary
evidence (reflexes of *parna- < *prno- and forms derived from
it).
The Hittite paradigm (Nom.sg. per/pir, Gen.
parnas, etc.; also parnant- 'household', parna 'home [adv.]', parnawa:i-
'build') is evidently archaic: *pe:r, *pr-n-ó-, not unlike *k^e:r,
*k^rd-(j-)ó- (> Hit. gir, kardijas 'heart'). The form *pe:r is often thought
to reflect *per-r (thus in the EIEC) or *per-n with compensatory length,
though is could equally well represent the bare root *per-. Morphologically,
the word looks archaic enough to be PIE ...
... but on the other hand, we have no
certain non-Anatolian cognates. Khotanese pira 'house' looks promising but is
isolated even within Indo-Iranian, and for lack of related forms we can't
reconstruct its protoform with sufficient precision to tell if it's comparable
with Anatiolian *per-/*pr-n-o- at all. The existence of Egyptian p.r- and
Hurrian purli 'house' suggests that *per- may be a Near-Eastern Wanderwort,
perhaps ultimately of Egyptian origin (the Hattic word for house,
<wel>, is of course unrelated). It would have had to be borrowed into
Anatolian rather early -- so early that it could pass for an inherited word
...
... but on the other hand, it has been
tentatively suggested that *prijo- 'dear, one's own, friendly' is derived from
*per- 'house, home'. Anyway, the verb stem *pri-ah2- 'love, prefer' is
de-adjectival (like *new-ah2- 'renew') and *pr-i-o- [prijó-] is a conceivable
extension of *per- ...
... but on the other hand, this is a
mere guess, the semantics is somewhat loose, and since PIE has no shortage
of terms for 'house(hold)', 'home', etc. (*dom- and *k^oi-mo-, for example), one
may legitimately wonder what semantic niche is left for *per-.
No definite conclusion.
Piotr