How to kill a cow in Avestan
175
person active forms,
avocam (1.sg.) and
avoc¯ama (1.pl.) that are normally used as in
RV 1,114,11a-d: ávoc¯ama námo asm¯a avasyáva ˙h “(with these words), we have
proclaimed the (hymn of) veneration, longing for help”
38
. The same principle can be
seen in the use of the Old Av. counterpart of
avoc¯ama attested in the finishing stanza
of Y. 38 in the form ¯auuaoc¯ama
39
which may conceal ¯a plus auuaoc¯ama: “(with these
words), we have called you hither (the waters ...)”.
40
Thus, the assumption that the 3rd sg. passive forms in -i pertain to the aorist system
in Indo-Iranian seems well founded. This does not mean, however, that they formed
a part of the aorist paradigm proper: There is a clear difference between (a)v¯aci on the
one hand and (a)vocam, (a)voc¯ama etc. on the other hand in that the latter only are
built from the reduplicated thematic stem (*e- ˘ue- ˘uk
˘u
e/o-) which by comparison with
Greek e{ipon can be regarded as inherited from the I.-E. protolanguage. (a)v¯aci,
however, must be considered as an athematic formation, consisting of an ending -i
directly attached to the o-graded root
41
. Taking this into account for the case of mrao¯ı
meaning “it is spoken”, I have argued (1985, 55 n. 65) that this could easily have
been built by analogy with the passive aorist forms, because
√mr¯u had an athematic
root present. The analogy would thus have consisted in transferring the ending only,
yielding /mra ˘ui/ with short -a- in accordance with the B
RUGMANN
condition of a
closed syllable produced by the root final laryngeal (*mro ˘u.H-i)
42
.
But such an assumption is not even necessary to justify mrao¯ı, given that within
Avestan, at least one passive form is attested that is regarded by the
communis opi-
nio
43
to be built from a marked present stem. This is e r e n¯auui which occurs several
times in the so-called H¯om Yašt (Y. 9) in the formula
k¯a. ahm¯ai. aˇ˙siš. e r e n¯auui.
“what an allotment was allotted to him?” (Y. 9,3.6.9.12; the following verses each
contain the answering formula h¯a. ahm¯ai. aˇ˙siš. e r e n¯auui. “this allotment was allotted
to him”). Note that e r e n¯auui and aˇ˙si- (< *árti-,
√
2
ar) are joint in a figura etymologica
which reoccurs, with plain medial forms functioning as passives, in Y. 56,3-4 and
65,17: .. va
o
huii˚¯asc¯a. aˇ˙s¯oiš. yasn¯ai. y¯a. n¯e. ¯ara¯ec¯a. e r e nauuata¯ec¯a. aˇ˙sa
o
h¯axš “.. for
worship(ping) of the good allotment which was allotted to us (formerly: pf. ¯ara¯e) and
will be allotted to us (in future times: cj. e r e nauuata¯e), accompanying (or accom-
panied by) truth”.
aorist indicative in general (“aktuelle Vergangenheit”, “resultative Konstatierung”).
38
Similarly: ávocam: 1,116,25a; 1,185,10a; 4,45,7a; ávoc¯ama: 1,78,5a; 189,8a; 4,2,20b; 5,1,12a;
5,73,10d; 10,80,7b; exceptional: 8,59,5a.
39
This variant, present in Pt4, K5, J2, S1; Mf2, Jp1, K4, is preferable as against G
ELDNER
’s
auuaoc¯ama taken from Dh1, Lb2, H1, L13, J7, P6 or auu¯oc¯am¯a as represented in Mf4, Mf1, J3, L2,
L1, O2, B2, L3, Bb1, C1; cf. N
ARTEN
1986, 235 n. 145.
40
For auu¯ac¯ı, the last word of Y. 36, a similar analysis is hardly possible; cf. below.
41
A recent attempt to find an I.-E. perspective for this formation was published in J
ASANOFF
1992, 129 sqq.; now cf. also K
ÜMMEL
1996.
42
Cp., e.g., Ved. (á)jáni < *(é)´gónh
1
-i from se ˙t
jan
i
“to beget” vs. át¯api < *étop-i from ani ˙t tap
“to heat”; H
UMBACH
’s *
mr¯auu¯ı (1991: II, 89) has no basis.
43
Cf., e.g., B
ARTHOLOMAE
1904, 184 f. s.v.
2
ar-; K
ELLENS
1984, 231; H
OFFMANN
/F
ORSSMAN
1996, 228.
176
Jost Gippert
A second Avestan form that has to be dealt with in this connection is jaini which
occurs three times in Yt. 19,92-93. This can only be analysed as a passive form
belonging to
√jan “to kill”: It appears just in the ritualized context of killing enemies
Calvert W
ATKINS
discussed in the article we started from. Cf. his interpretation of Yt.
19,92
44
:
+
vad
e m va¯ej¯o yim v¯ar e \
rag
n e m
yim bara ˜t taxm¯o \
ra¯etaon¯o
ya ˜t ažiš dah¯ak¯o jaini
‘swinging the weapon which smashes resistance
which brave Thraetaona carried,
when Aži Dah¯aka was slain.’
With the object of slaying, aži- dah¯aka-, put in the nominative, this verse represents
a clear passivization of the active clause present in Y. 9,8 ... \ra¯eta¯on¯o .. y¯o jana ˜t
až¯ım dah¯ak¯em ‘... Thraetaona ... who slew the dragon Aži Dah¯aka’
45
. Of course we
have to note that for the verbal root in question,
√jan < √g
wh
en-, we should expect not
jaini but *
j¯aini as the 3rd sg. passive form containing the
o-graded root, *
(e)g
wh
on-i —
just as we should expect not jana ˜t but *jan or *j e n (< *g
wh
en-t
≈ Ved. hán, Hitt.
kuenta
46
) for the 3rd sg. act.inj. form. But nevertheless, there can be no doubt that
jaini was built directly from a verbal root which shared two important features with
√mr¯u “to speak”: Both were typical “Präsenswurzeln” in the sense that they formed
an athematic root present and never developed an aorist stem of their own.
It is interesting, then, to see that in Vedic too, we find i-passives from roots that
form present stems primarily or exclusively. One such case is bh¯ari (RV 9,97,23d).
Although there are but few traces in Vedic showing that
√bhar “to bear” once formed
an athematic root present (3rd sg. pres.ind.act. bhárti in RV 1,173,6d, bharti in
6,13,3b; what we normally have is a thematic full-grade root present, bhárati, or a
reduplicated athematic one, bibhárti), and although an s-aorist of this root is attested,
it is quite probable on comparative grounds that
√bhar was inherited as a “Präsens-
wurzel” into Indo-Iranian
47
. One more such case is
√
stu “to praise” whose
i-passive,
á/ast¯avi, is attested six times in the RV (1,141,13a; 6,23,10b; 8,52,9a; 10,45,12a;
63,17d; 64,17d). Besides being notorious for representing the special “Narten” type of
athematic root present, this root too has an s-aorist in Vedic; but as with
√bhar, it
seems likely that this is only secondary
48
, all the more since it has no counterpart in
Avestan.
44
Cf. W
ATKINS
1987, 275 (where “Yt. 19.32” is a misprint). Although a variant vad e m seems not
to be attested in the present place, the emendation of va¯ed e m (v.ll. vaid e m, vaed e m, vaid e m: H
INTZE
1994, 370) as proposed by S
CHINDLER
apud W
ATKINS
l.c. is preferable to B
ARTHOLOMAE
’s view now
supported by H
INTZE
(o.c., 373 sq.) according to whom this might be a derivative of the root present
in OInd.
vídhyati “to shoot, to hit”. Cp.
vad e m attested in Vd. 14,7 (L4) with variants
vadim (Jp1),
vad
e m (L1, M2, O2), but also
va¯ed
e m (K1, K10, L2, Dh1),
vid
im (Mf2),
vaid
e m (B2), or the cognate
vadar¯e beside
v¯oižda ˜t in Y. 32,10c.
45
W
ATKINS
, o.c., 274.
46
Cp. Old Av. aj¯en in Y. 48,10 if this represents *¯aj¯en < *¯a-ˇ
i
h
ant as suggested by K
ELLENS
1984, 94, H
UMBACH
1991, II: 203 or H
OFFMANN
/ F
ORSSMAN
1996, 201 (B
ARTHOLOMAE
1904, 492
takes this as a 2.sg., quasi *¯a-ˇ
i
h
ans).
47
For the Vedic aorist forms cf. N
ARTEN
1964, 183 according to whom these “machen .. den
Eindruck .. einer Neubildung”; for further literature, cf. M
AYRHOFER
1986-, II: 248 sq.
48
Cf. N
ARTEN
1964, 276 sqq.