Origin of the Verbal Endings

The following system of verbal endings has been reconstructed for the latest stage of Proto-Indoeuropean:

ending category

primary active

secondary active

primary medium

secondary medium


first person singular -mi -m (-moi) (-mo) -h2e
second person singular -si -s -soi -so -th2e
third person singular -ti -t -toi -to -e
third person plural -nti -nt -ntoi -nto -r
marker present tense marker -i unmarked (original) form medium voice + present tense marker -o + -i medium voice marker -o special perfect endings, marker -e?

It is not less probable that the full ("primary") forms active and medium are the original ones and the secondary forms are reduced ones - this assumption leads to another interpretation of the semantic content of the endings.

I would suggest the following set of endings as the original one:

ending category

transitive verb

dative construction

intransitive verb

common interpretation "primary active" "primary medium" "perfect"
examples -mi, -si, -ti, ..., -nti -moi, -soi, -toi, ..., -ntoi -h2e, -th2e, -0e, ..., -r
reduplication vowel -i -i < -ei/-oi due to accent? -e
case of the agent ergative = "locative" reflexive = "dative" absolute = "accusative"
case ending (singular) -i -ei/-oi -e (personal pronouns) or -0 (nouns) + -m/-d (gender)

This system leads to the following conclusions:

Written by Hans-Joachim Alscher