Re: [tied] a

> >
> > ***
> >
> > > Under certain conditions (no stress, stress, unknown),
> > > /a/ > nothing, /e/, /o/
> > >
> > > Under the same set of conditions
> > > /i/ > /i/, /ei/, /oi/
> > >
> > > And, under the same set of conditions
> > > /u/ > /u/, /eu/, /ou/
> >
> >
> >
> > > None of this is phonetically implausible. Re the reflexes

of /i/

> > > and /u/, cf the inflection of nominal i- and u- stems.
> > >
> > > We now have something that is identical to the classical view

of

> > > PIE, except that the striking similarity between the three
> series of
> > > reflexes is spurious (but nonetheless real). > > >
> > > The problem of this analysis (ie starting from the zero grade)

is

> > > how to account for i- and u- ablaut in the other direction,
> > > namely /i/, /ie/, /io/ and /u/, /ue/, /uo/
> >
> > ***
> > Patrick:
> >
> > Could you elucidate what you mean here?
> >
> >
>
> You can either choose full grade (/eu/ vs /ue/ as your starting
> point, in which case you easily separate /u/, /eu/, /ou/
> from /u/, /ue/, /uo/, or zero grade /w/ as I do, in which case you
> will have trouble deriving both /eu/, /ou/ and /ue/, /uo/

from /u/;

> what is the rule which selects one or the other?
>

Jens in 'Szereményi's Theory of Indo-European i- and u-Stems'
mention Szereményi's distinction wrt. IE i- and u- stems of a type I
with *-yo-s, *-wo-s in the weak cases like gen.sg., and a type II
with *-ey-s, *-ew-s (might be *-oy-s, *-ow-s) (I've re-analysed
them). But that means that one can get (to quote myself):

Under certain conditions (no stress, stress, unknown),
/a/ > nothing, /e/, /o/

Under the same set of conditions
/i/ > /i/, /ey/, /oy/ _and_
/i/ > /i/, /ye/, /yo/

And, under the same set of conditions
/u/ > /u/, /ew/, /ow/ _and_
/u/ > /u/, /wu/, /wo/

Now if this new functional load on (zero, e, o), abstracted from the
spurious similarity of (zero, e, o / i, ey, oy and i, ye, yo / u,
ew, ow and u, we, wo) would spread to other instances, then we can
go from a three-vowel pre-PIE system (i, a, u) with no functional
load (ablaut) to a five-vowel ablauting system of classical PIE
_without_ passing an functionally unlikely single-vowel (A or schwa)
system that everyone else proposes to account for the origin of
ablaut.


Torsten