On 2007-06-11 21:16, Sean Whalen wrote:
> Forms with and without h1 appear independent of
> placement in compounds; it seems to me that it's a
> stative morpheme.
Forms without a laryngeal appear very consistently in compounds and
reduplications; only rarely elsewhere, and those "elsewhere" variants
may be decompositional. Examples are so familiar that they scarcely need
to be quoted: beside *newo-g^n-o- (with the exact same treatment in
several branches far apart) and *kWe-kWl-o- we have e.g. Skt. perf.
ja-jn~-é ("without" *h3) and <pra-tn-á-> ("without" *h2), numerous
compounds with *-pl-o- (from *pleh1-), etc. Of course compositional
shortening is a strong cross-linguistic tendency (cf. <twopence>, the
traditional pronunciation of <forhead>, etc.), so no wonder that in some
cases it's hard to say whether a given instance of it is PIE or
branch-specific.
Piotr