How weird is Hittite? Not weird enough :)

Miguel Carrasquer Vidal mcv at
Thu Feb 4 19:37:33 UTC 1999

Vidhyanath Rao <vidynath at> wrote:

>I don't find it obvious that the secondary endings expressed tense
>originally: The PE are SE extended with -i (meaning ``here and now''?).
>The imperative looks like just SE (in 2nd pl) or SE extended with -u
>(3rd), except in 2nd sing. Optative is built on SE. If SE expressed
>tense, how did these evolve? In particular, present = past + ``here and
>now'' looks strange.

Exactly, the fact that the imperative has the secondary (short)
endings shows that these were the unmarked, neutral ones.
Present (Non-past) = neutral + ``here and now''.  The forms
without the -i extension then become past forms (aorist or
imperfect) by default.  But the distinction was already in
Anatolian (-mi present vs. -m past).

>I also find Szeremenyi's objections to the view of PIE in any stage had
>morphologically expressed aspect convincing and think that they apply to
>any stage that includes both Greek and Indo-Iranian . In particular I am
>not sure that Greek and Indo-Iranian show similar verb systems.

The Greek and Indo-Iranian verbal systems are not identical of
course, but they are very very similar.  Much more similar to
each other than to any other IE verbal system.

Some features that are unique to the "Indo-Greek" verbal system

- the perfect as a separate "aspect", besides present (impfv.)
and aorist (pfv.).  Balto-Slavic has lost practically all traces
of the perfect (only the praeterito-preasens OCS ve^de^ "I
know"), so we might hypothesize that it existed in
Pre-Balto-Slavic.  Elsewhere, the perfect has merged with the
aorist (I'm assuming that at least some of the Armenian aorist
personal endings are derivable from the perfect).  In Hittite,
the perfect is still simply the past tense of the stative (hi)

- the imperfect as a simple past tense of the present ((augment
+) present stem + secondary endings).  The Albanian j/n-imperfect
and the Baltic j-imperfects are similar, except that they always
include an explicit imperfective marker (j, n) even if the
present tense forms don't.  Elswhere we find a variety of
different formations of the imperfect (based on the optative in
Tocharian and Armenian, sigmatic forms in Slavic and the Albanian
sh-imperfect, forms with *a: in Latin, Irish, Baltic, etc.
Hittite and Germanic (unless the weak *dh forms were originally
imperfects) have no imperfect at all.

- the subjunctive (conjunctive) as a thematic (of athematic
verbs) or doubly thematic (of thematic verbs) formation, without
additional markers.  The only parallel is I think Latin ero:, the
future tense of "to be".

- The augment for past tenses.  Also found in Armenian (
of monosylabic verbs only).

Additionally, another important feature of Greek and
Indo-Iranian, not shared by some of the other languages, are the
sigmatic aorist and future forms, which are probably related (as
in the Slavic languages, the present tense of the perfective
makes the future).  If so, in Skt. the s-marker of the future is
combined with imperfective/presentive *j, giving -sya-.  We find
sigmatic futures (subjunctives) also in Albanian (=optative),
Baltic, Celtic, and sigmatic aorists in Slavic, maybe in
Albanian, and mixed up with perfect forms in Latin, Old Irish and
maybe Tocharian.  Armenian has *(i)sk (-(i)c`-) instead
(subjunctive and aorist).  Hittite and Germanic lack sigmatic
forms (and I have my doubts about Tocharian).

Contrary to the traditional view, in which Hittite and Germanic
have lost categories such as the s-aorist and the imperfect, I
think that Hittite is closest to the original state of affairs.
The present tense of the mi- and hi-conjugations gave the
athematic (*-mi) and thematic (*-o-H2) present tenses (with
considerable mix-up of the two, as indeed in Hittite itself).
The past tense of the mi-conjugation (the m-past) was preserved
in the Indo-Greek imperfect and root aorist, but the general
tendency was to replace it with extended forms (sigmatic aorist,
the different "innovative" imperfects).  The past tense of the
hi-conjugation, with its characteristic endings (*-H2e, *-tH2e,
*-e/*-s), became the perfect in Indo-Greek (and maybe once
Balto-Slavic?), the unmarked past tense (preterite) elsewhere.
Additional markers were often added (reduplication [if not
original], the Greek k-perfect, the w-perfect, etc.)

In summary, I'd say that the Greek and Indo-Iranian verbal
systems have many things in common and are closest to the fully
developed "Brugmannian" model [which contains some features which
are archaisms, such as the simple "m-past" (imperfect/root
aorist)].  Despite the fact that Balto-Slavic has developed new
perfective/imperfective categories (using preverbs) which have
eliminated the old system (e.g. aorist and imperfect) in most of
the modern languages (incl. all the Baltic ones), Balto-Slavic
seems to come closest to the Indo-Greek system, especially the
OCS root and sigmatic aorists ("uncontaminated" by perfect
forms).  Italic, Celtic and Albanian also share the "sigmatic"
isogloss with Greek, Indo-Iranian and Balto-Slavic, but their
forms are best described as s-preterites.  Additionally Italic,
Celtic and (possibly) Albanian are characterized by the use of
secondary endings in the present tense (loss of functionality of
the primary-secondary distinction?).  The Armenian verbal system,
despite the presence of the augment, seems to have little in
common with the "Indo-Greek" one.  The use of *sk in the aorist
and subjunctive cannot be simply equated with the sigmatic forms
we find elsewhere.  And if the Armenian imperfect is indeed
derived from the optative, that's a remarkable Armenian-Tocharian
(and Italo-Celtic?) isogloss.  The most archaic verbal systems
(or at least the ones least similar to Greek and Indo-Iranian)
are those of Tocharian, Germanic and especially Hittite.

If we take the active past tense forms with *s as our primary
isogloss, we have:

1) no -s or *-s is a personal marker: Hittite, Tocharian, Germanic.
3) *-(i)sk-: Armenian.
4) s-preterite: Italic, Celtic, Albanian.
5) s-aorist: Greek, Indo-Iranian, Slavic(-Baltic).

Miguel Carrasquer Vidal
mcv at

More information about the Indo-european mailing list