12. The imperative
The imperative endings in the active (singular strong, plural/dual weak stem)
can be reconstructed as follows:
2. -0
3. -tu
2. -té(r)
3. -éntu
2. [-táh2]
3. [-áh2tu]
In the 3rd. person forms, we have a suffix *-u, which behaves similarly to the
present tense affix *-i. In particular, it does not affect the Ablaut: the 3sg.
imper. has the accent on the stem.
The dual forms as actually attested are:
Skt. Grk. Goth
2 -tam -ton -ts
3 -ta:m -ton
In other words, these are the normal indicative endings (in Indo-Iranian the
secondary ones, in Greek the primary ones; Gothic does not distinguish the two).
If the dual ever had specific imperative endings, they have been lost.
The 2sg. has a zero ending, which means that in the thematic verbs the 2sg.
imperative ends in *-e. Athematic verbs can also take the zero ending, but
often a particle *-dhí is added, which attracts the accent and leaves the stem
in zero grade. The origin of *-dhi is unclear. Perhaps it comes from the
stative, where an ending **-tku should have given *-dh(w), as in the 2pl.
stative and middle.
Besides the normal imperative, there is a "future imperative", which has the
ending *-to:d in all persons except the 3pl., which has *-nto:d. This is
generally considered to be an ablative of a verbal noun in *-tu (identical in
form with the 3sg. imperative in *-tu, and therefore associated with the
imperative?). Gothic (-dau) and perhaps Old Irish (-th) have *-to(:)u instead,
which could be the dative of the same verbal noun in *-tu.
The attested middle imperative forms are:
Skt. Grk. Hitt. OIr.
2 -sva -so -hut(i) -the < *-te:C
3 -(t)a:m -stho: -(t)aru -th < *-tO
2 -dhvam -sthe -dumat(i) -th^ < *-tE
3 -nta:m -stho:n -antaru -tar < *-ntOr(V)
2 -a:tha:m -sthon -- --
3 -a:ta:m -sthon -- --
It is difficult to discern a common pattern here. Greek has the middle
secondary (injunctive) endings, except in the 3rd person sg. and pl., which have
analogically reshaped forms based on *-to:d/*-nto:d, and the 3rd dual, which has
the primary ending, as in the active imperative. Sanskrit has the peculiar 2sg.
ending -sva, which is probably agglutinated *swe "(your)self", as befits the
reflexive semantics of the middle. The 2pl. and du. forms are the same as the
injunctive endings. The 3rd person sg. and pl. seem to be middle forms with
1du. dative agreement (*-(t)ah2m, *-ntah2m).
The Hittite 2sg. ending -hut(i) is remarkable in having an element *-h2- which
is nowhere else attested as a second person ending in Indo-European. We would
have expected a reflex of *-th2, or possibly *-dhw (if that's the origin of
athematic active *-dhi), ultimately from the stative 2nd person ending **-tkV.
As it is, the Hittite form points to **-kV (cf. the Uralic 2sg. imper. ending
*-k) . The other Hittite forms show normal (indicative) middle endings,
extended with the imperative marker *-u.
The Old Irish middle imperative is unremarkable (injunctive endings), except for
the 2sg. ending -the < *-te:C. A connection with the deponent ending -ther has
been proposed, and if that form is a generalized absolute ending (*-th2ar+es),
then perhaps the imperative (which otherwise lacks absolute forms) is too:
*-th2a+es may have given *-te:s > -the.
=======================
Miguel Carrasquer Vidal
mcv@...