From: Piotr Gasiorowski
----- Original Message -----
From: "g" <gs001ns@...>
Sent: Thursday, January 02, 2003 12:06 PM
Subject: [tied] Re: -poulos and -putra
> >I forgot to add that <pullus> is ultimately connected with ...
> And with Lat. "puer"?
Yes, presumably. Also with Gk. pa(w)is. There is a tendency (visible in the American Heritage Dioctionary entry I referred to) to lump all those *p(a)u- terms together under a single "root with extended variants", one of the latter being *pu-t(o)-. I'm inclined to accept the analysis of *pullus as *put-lo-s, based on Nyman's observation that in root-suffix sandhi *tl gives /ll/ rather than /k(u)l/ (as in the suffix -c(u)lum < *-tlom) in Latin. See also Lat. putillus 'birdie, nestlig', which Szemerényi analyses as a diminutive of *putlo-. Since we also have Oscan puklo- 'son' to match both Indo-Iranian *putra- and Lat. pullus (cf. Osc akno vs. Lat. annus < *atno-), I see no reason to separate these words despite the varying semantics (the gap between 'child' and 'little bird' is not unbridgeable; even Slavic *pUt(ic)a might be added here).