--- In
cybalist@yahoogroups.com, "alex" <alxmoeller@...> wrote:
> petusek wrote:
> >
> > So, just as Richard says, it is not only unusual to take place
in a
> > single step, but also improbable and, as far as I can remember,
> > unattested.
> >
> > Petusek
>
>
> I remember about Latin loosing its initial velar.
>
> haedera < *ghed
> heri: < *g^hes-
> habeo: < *ghabh-
> haedus < *ghaidos
> etc
To be compared with *bH- > f-, *dH- > *รพ- > f-, *gHW- > *xw- > f-.
This change presumably went *gH > *x- > *h-. It's a bit misleading
to call this the loss of an initial velar, though the ultimate
process *gH > 0 is of course a plain loss, and the loss had sometime
happened by Classical Latin, as in *gHans- > _anser_ 'goose'. And
Sanskrit _does_ lose the velar element here - Sanskrit
_ha.msa_ 'swan'.
What you (Alex) probably had in mind was the example of the loss of
the velar element - *gW- > w-, as in *gWih3wos > Latin
_vi:vus_ 'live' > Spanish _vivo_ (cognate with English _quick_ and
Greek _bios_), *gWam- > Latin _venio:_ (as PIE "*gvmye'", one of
Sean Whalen's examples last weekend; cognate with English _come_) >
Spanish _venir_. Thus in the development from PIE to Spanish we do
have the regular word-initial change gW > w > b as Alex was
suggesting. By contrast, in Spanish we have PIE *kW > k ! My
issues were that (1) gW > w > b is not the route take in Latin or
Greek (or for that matter Welsh) and (2) kW > w (or W) > p does not
work in Italic languages, ancient or modern (i.e. Romance).
Richard.