I was reading the old messages in this forum and I found a very
interesting discussion that took place many months ago. It has
inspired me to offer my own views about Pre-PIE verbal morphology
(although they are not entirely original):

The 1st and 2nd person endings arose from enclitic pronouns (*me and
*te, respectively). Thus, with the root *bher-:

bher-me 'I bear'
bher-te 'you bear'

Later on, a distinction between definite (direct object overtly
marked) and indefinite arose (which quickly evolved into a transitive-
intransitive dichotomy). Pre-PIE speakers formed a definite
conjugation by adding the deictic/pronominal element -e- to the verb
stem. Thus:

bher-e-m(e) 'I bear it'
bher-e-t(e) 'you bear it'
bher-e '[it] bears it'

The 1st and 2nd plural endings were formed by adding the plural
suffix -es:

bher-m(e)-es 'we bear'
bher-t(e)-es 'y'all bear'

bher-e-m-es 'we bear it'
bher-e-t-es 'y'all bear it'

When distinction between singular and plural numbers in verbs
developed, the 3rd person forms were extended by the pronominal
element -t in the singular and the participial suffix -nt in the
plural:

bher-t 'it bears'
bher-nt 'they bear'

bher-e-t 'it bears it'
bher-e-nt 'they bear it'

At this point, I am unsure when and how the 2nd singular ending
changed to -s. Perhaps when combined with the non-past tense suffix -
i, the combination -t-i > -s-i by assibilation, and then spread
analogically throughout the rest of the paradigm(s).

More on tense:

At some stage in the above development, tense distinctions arose.
Two primarily elements were used: the deictic element -e- was
prefixed to the verb stem with a temporal meaning (*'at that time'

> 'then'); the locative marker -i was suffixed to the singular

definite/transitive endings to keep them distinct from the
indefinite/intransitive ones, and had the temporal meaning of *'hic
et nunc.'

As others have pointed out (Sihler), the PIE 'perfect' paradigm is
really a stative paradigm. I think the stative forms were always
created by reduplication of the verb root, even in very ancient
statives like *woid- (from **we(i)-weid-).

That's all for right now. Any thoughts?

- Rob