>What is the derivation of the Latin perfect endings? It seems to me as if
they are a mix of the sigmatic aorist and perfect endings.
They are more perfect than aorist. The PIE endings of the perfect appear
to have been:
-a = -h2e
-tha = -th2e
-e
me
?
-r(o)
Latin appears to have added an -i to these in the singular, and 3 plural,
just as -i was added to other primary endings in PIE (secondary
*-m, -s, -t, -> primary *-mi, -si, -ti). The third person singular then
picks up a final -t by analogy. The earliest attested Latin forms are
- ai
-is-tai
-eit
-e:ri
The 1 & 2 plural pick up the -s under analogy, as elsewhere in the verbal
system. The 2 plural appears to be the 2 singular with -s added.
An original aorist 3 singular should have given early Latin -ed (as in the
Praenestine fibula, the authenticity of which is sometimes disputed).
The -is- element is not easily explained, but probably it is not the
sigmatic aorist. The sigmatic aorist forms would appear as stem + s (not
stem + pefect extension + is). These forms are actually found: amasti,
dixti, and so on. Some are later contractions of the perfect stem, but some
cannot be (eg the forms like putasti).
The 3 plural shows both forms: -e:ri > -e:re and -is-ont > erunt (short
e). These are merged to the classical form -e:runt.
Thus there is little, if anything, of the aorist in the Latin perfect
endings.
Peter