Miguel:
> If we take PIE -i as a locative ending and translate it as
> "in", the present tense forms in -i may be taken as
> parallel. Except that the -i is appended to the personal
> ending, not to the verb (c.q. verbal noun), which is odd.
I've already explained my view that the indicative *-i is nothing
more than yet another demonstrative being fused to preexisting
forms. So we can think of *bHereti as originating quite simply
from a _phrase_. In Mid IE, it would look like this:
*béra-ta ei
literally "carry-this here/now"
"he/she/it is carrying"
Here *ei would mean not so much "at this place, here" but "at this
time, now" or to emphasize the reality of the action. So I think that
the indicative is an agglutinated phrase using a locative particle,
completely parallel with a mediopassive *-r originating from an
agentive particle.
After Syncope, MIE *béra-ta shrinks regularly to eLIE *bHérat and
the locative particle *ei becomes *i, yielding a new indicative
*bHérat-i. Since unstressed *a becomes *e before unvoiced consonants,
we get the expected pair without problems, *bHereti/*bHeret.
And the rest, as they say, is prehistory.
= gLeN