3. The preterite
Under the heading "preterite" I will discuss those forms which are usually
called imperfects, root aorists and thematic aorists in Greek/Sanskrit and
Indo-European grammars. Sigmatic aorists (thematic and athematic) and perfects
will be treated elsewhere.
The active preterite endings ("secondary endings") are like the active present
ones, but without the *-i. The alternative set of endings has *-é instead of
present *-és(W) (perhaps influenced by the perfect endings, for which see
later).
Set I:
**-mW **-mwén **-mwáh2
**-s(W) **-t(w)ér **-t(w)áh2
**-t **-ér-[t] [**-(y)éh2-t]
Set II:
**-mwé **-mWh2á
**-t(W)é **-t(W)h2á
**-(h2)té
Note the regular development **-n[C] > *-r[C] in the 2nd and 3rd plural, but not
in the 1pl., where *-n was preserved after nasal *-m-.
Hittite reflects this as:
athematic thematic
-un -men -a-nun -a-men
-s -ten -e-s -e-tten
-t(a) -er, -ir -e-t -er, -ir
In the 1st sg., -un is the regular reflex of syllabic *m., the form -nun (*-m +
*m.) is used after vowels (e.g. in the thematic forms). The 2nd plural form is
analogical for expected *-ter.
Tocharian B has:
athematic thematic
-m -(ä)m(o) -em -em(o)
-(ä)t(o) -(ä)cer -^t(o) -^(ä)cer
-0 -(ä)m. -^ -em.
The 2nd sg. form has the characteristic Tocharian -t. The 3rd sg. form lacks
present tense -m. (< *nu), as expected. The plural forms can be derived from
*-me or *-men, *-te:r and *-ent(i). The form -cer (< *-te:r) is interesting, as
it's the only indirect attestation of expected 2pl. *-ter (except if Skt. 2du.
perfect -athur is a transfer from the plural). The long vowel can be explained
as analogical after the 3rd plural form *-e:r (< *-er-s), which was later
replaced itself by analogical *-ent in Tocharian (and throughout most of the IE
area). Tocharian B uses these past endings in the optative and the imperfect.
The present tense forms are identical, except that 1sg. has -ew (= thematic
*-o:u), and the 3sg. adds -m.. Tocharian A uses the present endings in both
present and imperfect/optative. The Tocharian preterite has endings which are
derived from the PIE perfect (see below).
The Germanic preterite is derived from the perfect (see there). West Germanic,
however, does have an ending *-i in the 2sg. preterite, which is added to the
plural root (zero grade in most verbal categories). It might be derived from a
2nd. person preterit (thematic aorist) in *-és (> *-iz > *-i), although the
rationale for such a replacement remains mysterious.
The endings of the Armenian preterite cannot be derived from the PIE simple
preterite (asigmatic aorist or imperfect). I will treat them later, under the
perfect.
In Old Irish, only the singular of the t-preterite, made from strong verbs in
mainly -l and -r, continues the PIE preterite (root aorist), albeit in a
peculiar way:
.biurt
.birt
.bert
The basic form was here the 3 sg. root aorist *bher-t "he carried", where *-t is
of course the 3rd. person sg. secondary ending. Taking *bhert as the past tense
stem, the 1 and 2 sg. were reshaped to *berto:, *berti: (the endings are those
of the thematic present), resulting in the Old Irish paradigm as given above.
The Latin preterite continues, at least in its endings, the PIE perfect.
The Lithuanian preterite is made with the suffixes *-e:- and *-a:-, on which see
below.
In OCS, the simple aorist is conjugated as follows:
-U -omU -ove^
-e -ete -eta
-e -o~ -ete
This reflects the PIE thematic aorist (*-om, *-es, *-et; *-omos, *-etes, *-ont;
*-owa:, *-eta:, *-ete). The 2nd. and 3rd. person sg. forms of this paradigm are
also used in most verbs to substitute for the 2nd and 3rd persons of the
sigmatic aorist (where the endings *-s(s), *-st had been reduced to zero).
The Greek imperfect and root aorist endings:
athematic thematic
-n (-a) -men/-mes -- -on -omen/-omes --
-s (-as) -te -ton -es -ete -eton
-0 (-e) -san -te:n -e -on -ete:n
The 3pl. -san (*-s-n.t) is borrowed from the sigmatic aorist. The dual forms
are originally middle duals (see there).
Vedic imperfects and root aorists:
athematic thematic
-am -ma -va -a-m -a:-ma -a:-va
-s -ta(na) -tam -a-s -a-ta -a-tam
-t -ur,-an -ta:m -a-t -a-n -a-ta:m
The 3rd plural athematic has -ur (< *-rs, Avestan -rs^) or -an (-ent, -n.t), the
thematic forms always *-ont (expected thematic **-or[t] is not found anywhere:
only the analogical form *-ont is attested). As in Greek, the 2/3 du. forms are
old middle duals.
In Greek and Sanskrit, a distinction is made between imperfects, where the
secondary endings are added to the same stem as the present, and (asigmatic)
aorists, where these same endings are attached to the aorist stem. The
distinction between present and aorist stems is a subtle one. A simple root
with secondary endings is an (asigmatic) aorist if the verb in question has an
extended present stem (made with -sk^e-, -ne- or -ye-, for instance), but it's
considered an imperfect if the root lacks such an extended present stem (in
which case the aorist is likely to be either a sigmatic aorist or a root aorist
made from a suppletive stem). This follows from the fact that some verbs by
their semantics are inherently imperfective (they denote actions which by their
nature are repetitive or do not cause results, e.g. "to sleep"), while other
verbs are inherently perfective. PIE had a number of suffixes that could be
used to turn perfective verbs into imperfective/durative/iterative ones, and
these devices led, but probably only after the breakup of PIE, to the systematic
present stem ~ aorist stem distinction that we see in e.g. Greek.
Against amalgamating the imperfect and the aorist at the PIE stage is the fact
that there is one important formal difference between the thematic aorist and
the thematic imperfect. The latter is conjugated like a normal thematic verb
(*[h1e-]bhér-et), while the former invariably follows the tudáti-type, i.e. the
thematic vowel is stressed and the root appears in zero grade
(*[h1e-]bhugh-é-t). The same is also the case with the thematic s-aorist (with
accented *-sé-). Why this is so is not clear. It does not appear to be an
exclusively aoristic thing: the aorist conjunctives (as in fact all
conjunctives) are root-accented, while most verbs with thematic "presentive"
extensions (*-sk^é, *-yé-) follow the tudáti pattern as well. Perhaps
originally all non-conjunctive thematics were accented on the thematic vowel,
and the root-accented type spread from the conjunctive to the present/imperfect
indicative, while leaving the aorist/preterite as it was.
In any case, in the 3rd person sg., if the accent was on the thematic vowel, the
*-t must have been affected by the post-Anatolian development *-ét > *-éh1 (cf.
the instrumental sg. Hitt. -it, elsewhere *-éh1). As we have seen, in Greek and
Indo-Iranian the ending was analogically restored based on the present (*-ti,
*-eti) and the root-stressed preterite (*-t, *-et), but elsewhere, 3sg. *-e: was
the starting point for a new preterite formation (in much the same way as
athematic *-t was the starting point for the Old Irish t-preterite). This
e:-preterite is characteristic of Lithuanian, where we have:
-iau (< *-e:u < *-e:-o:) -e:me (< *-e:-me:)
-ei (< *-e:i < *-e:-ei) -e:te (< *-e:-te:)
-e: (< *-e:[t]) --
The type must also have existed in Slavic, where it perhaps survives as the
e:-element in the imperfect -e:-axU, etc., on which more later. If so, then the
e:-element of the Latin imperfect, which, besides unsurprising -e:-bam in the
e:-stems, is also added to i-stems and consonant-stems (audi-e:-bam, em-e:-bam)
is likely to have the same origin. The e:-preterite also surfaces in Tocharian,
where it is especially characteristic of the preterite (technically, an
imperfect rather than an aorist) of verbs in *-sk^é- (pret. *-sk^é-t > *-sk^é: >
-s.s.a:). With unpalatalizable consonants, one cannot tell whether Tocharian has
an e:-preterite or an a:-preterite (-e: becomes -(^)a: in the Auslaut). In
Lithuanian too, the e:-preterite is in complementary distribution with the
a:-preterite. As we shall see when discussing the stative conjunctive, another
possible source of the e:-preterite is closely linked to the a:-preterite/
a:-conjunctive of Tocharian, Armenian, Celtic, Latin and Balto-Slavic, and we'll
defer discussion of such forms until that chapter.
=======================
Miguel Carrasquer Vidal
mcv@...