Re: Prenasalization, not ejectives cause of Winter's law?

> Let me remind you of your ingenious emendation of *gWih3w-e/o-
> "live" to a reduplicated *gWigW-e/o-. You ascribe that to
> dissimilation, but note that the *gWigW- part violates the root
> constraint against roots of the form *DeD-. This is interesting,
> since it might give an insight into how PIE dealt with
> 'perpetrators' which broke the rule, whether they were loan or
> composition. In other words, from /gW/ you might get /w/ and /h3/.
> So was it once *stergW- that was hit by something? Does Lat.
> stercus "manure" have something to do with it?

So (*NgW is my version of standard *gW, pre-nasalized):

PPIE *stárNgW-ti, *starNgW-ánti ->
PIE *stérgW-ti (-> *sterk-ti), *sterngW-énti (Hittite-fashion) ->
PIE *stern-ógW-ti, *sterngW-énti

perhaps *-NgW.- -> -u-/-uk- ?, cf.
struo: <- *str.NgW.-ó:-
structus <- *str.NgW.-tó-


Torsten