How to kill a cow in Avestan
173
complete wording is
〈
c
štyh w hmwndyh ycwm *wt
c
w
c
h wnkyl
c
y hmyst
c
l
〉, i.e. ¯ašt¯ıh ud
am¯awand¯ıh yazom *wad-xw¯ah *ud *nakk¯ır¯a ham¯est¯ar in K20, “J1”, TD23),
wadxw¯ah
and nakk¯ır¯a do not necessarily witness to the meaning of the Avestan words they
represent; as often, they may have been inserted by the translator as mere stand-ins,
suggested to him by the context. In the present case, we cannot but suspect that either
¯ašt¯ıh or
am¯awand¯ıh (or even
ham¯est¯ar) was used in comparable juridic environments.
It must be underlined, however, that by now, the Pahlavi translation can only be
quoted from mss. belonging to the K20-group; F1 contains a s¯ade text (as well as
R115), and for the other mss., it is not clear from G
ELDNER
’s description whether they
have Pahlavi versions or not. It seems that the F1-tradition is restricted to
s¯ade mss.
at all.
In connection with the other alleged cognates of the Avestan root
2
mr¯u “to violate”,
the Pahlavi tradition gains more weight. mr¯ur¯o, epithet of the winter in Vd. 2,22, is
translated by m¯udag which in its turn is glossed by
〈AYK MNDOM tp
c
h
O ¯BYDWNyt
〉, i.e k¯u tis tab¯ah kun¯ed “which makes thing(s) spoilt”
32
. The same
word,
m¯udag, is used as the equivalent of
mruta in the Farhang-¯ı ¯o¯ım (F. XI: 491). As
G. K
LINGENSCHMITT
(1968a, 149) points out, we find two further attestations of both
m¯udag and its gloss in the Pahlavi version of the H¯om Yašt, viz. in Y. 9,32 and 11,6,
where the text has the compound
〈mwtk-krt
c
l
〉, i.e.
m¯udag-kard¯ar “spoilt-maker”. The
former testimony is worthless because here, the compound obviously mirrors Av.
maod
an¯o.kairii¯a, a bahuvrihi meaning “whose action is lust” (attribute of a whore,
jahik¯a-);
m¯udag may have been chosen in this case because of its phonetic similarity,
the Av. hapax maodana- reminding of MPers. m¯ud-. In Yt. 11,6, however, m¯udag-
kard¯ar translates Av.
m¯urak¯a (nom.pl.), one in a series of three names of (Daevic)
creatures (beside dah¯ak¯a and var e šn¯a) to be born in the house of somebody who
deprives Haoma of his legitimate share. The exact meaning of the Av. word remains
unclear, although it might be a derivational form of m¯ura- as appearing in Yt. 5,93, a
hapax mostly translated as “wicked” in agreement with Ved. m¯urá-. Unfortunately, a
Pahlavi version of Yt. 5 is not available so that we cannot prove that the Pahlavi
translator had this word in mind when rendering m¯uraka- by m¯udag kard¯ar. There is
a hidden indication, however, of the interdependency of m¯ura- and m¯udag to be found
in the Farhang-¯ı ¯o¯ım. The entry following the hapax mruta is unusual in the sense that
the Av. word in question, m e r e z¯an¯ai, seems to be glossed by two Pahlavi lemmas, one
of them written in Avestan script. According to K
LINGENSCHMITT
(1968a, 1650), the
entry reads m e r e z¯an¯ai:
+
mwl
c
n cygwn
+
KLSH, i.e. “
+
mul¯an soviel wie Bauch” (KLSH
is the aramaeogram for aškam(b) “belly”). While Mpers. mul¯an “belly” and Av.
m e r e z¯ana-, a hapax again, can easily be identified etymologically (presupposing
OPers. *m e rd¯ana-), there is no reason why the former should have been written in
Avestan script, all the more since the writing is defective: what we read is mur¯a in
both substantial manuscripts containing the Farhang
33
. We may therefore suggest that
two items of the prototype manuscript were confused here, viz. (MPers.)
mul¯an and
32
Cf. the edition S
ANJANA
1895, 19; J
AMASP
/ G
ANDEVIA
1907, 36 have kun¯end “they make”.
33
Cf. K
LINGENSCHMITT
1968a, I and 1968b, I. Besides K20 and M51, the Farhang is included in
TD28 (ed. J
AMASP
A
SA
/ N
AWABI
1976i); here the text is as defective for the entry in question (159)
as the one contained in M51, reading only (m e r e z¯an¯ai mur¯a plus)
〈SE〉 instead of 〈ˇcygwn KLSE〉.
174
Jost Gippert
(Av.) m˘¯ ur¯a, the latter representing the proper equivalent (and etymological cognate)
of MPers. m¯udag (and mruta, a secondary marginal or interlinear gloss
34
).
Returning to mr¯ur¯o in Vd. 2,22, we may wonder then whether this might have
secondarily replaced an original *
m¯ur¯o meaning “noxious”. In this case, we should
have to explain the first r in mr¯ur¯o as an intrusion. A possible source for this can be
traced in Vd. 7,27 where, in another lamentation about bad winter, this is called xr¯uta-
“cruel” (and, n.b., gaojan- “cow killing”). The parallelism of xr¯ura-
35
/ xr¯uta- and
mr¯ura- / mruta- is indeed striking. After all, the adaptation of
*m¯ura- to
mr¯ura- need
not necessarily have been restricted to manuscript tradition. It may well have become
a feature of the spoken (Young-)Avestan language
36
. But even if it did, this does not
mean that we should expect the resulting “root”
√mr¯u to have been able to form a
passive root aorist, /mra ˘ui/, or a feminine ¯ı-stem noun mrao¯ı- in Gathic times — just
as *kreuh
2
-, the “root” underlying Av.
xr¯ura-, did not develop a primary verbal
paradigm in any I.-E. language.
What, then, is mrao¯ı in Y. 32,14? The main argument put forward by H
UMBACH
and K
ELLENS
when rejecting the traditional analysis of
mrao¯ı as a 3.sg. passive form
of
√mr¯u “to speak” was that this root formed a present stem only, its aorist being
supplied by
√v¯ac, and that the expected 3rd sg. passive forms, v¯ac¯ı (inj.) and auu¯ac¯ı
(ind.) are well attested (Y. 43,13 / Y. 36,6: H
UMBACH
1991, 89). It is indeed true that
the passive formation with the ending -i is confined to the aorist in Vedic. This can
easily be shown by looking at the attestations of the immediate cognate of Av. auu¯ac¯ı,
viz. Ved. av¯aci: In five of its six occurrences in the RV, it appears in the last stanza
of a hymn, its function consisting in stating that the aim of proclaiming the hymn has
just been achieved by reaching its end; cp., e.g., RV 6,34,5b: índr¯aya stotrám matíbhir
av¯aci “(with these words,) a praise song has now been proclaimed to Indra, with
(pious) thoughts”
37
. Thus, av¯aci is in perfect agreement with the corresponding 1st
34
Note that in K20, the Mpers. equivalent of mruta is written
〈mwltk〉 (cf. C
HRISTENSEN
1931,
fol. 82v). This suggests a (secondary?) identification of mruta with m e r e ta- “dead”
≈ Npers.
murda
< Mpers. *
murdag; cp. also n. 36 below.
35
xr¯urahe instead of xr¯utahe is the variant reading of Jp1, Mf2, K10, L2 in Vd. 7,27; other
variants are xratahe (K1) and xr¯uvahe (! L1, O2). G
ELDNER
’s
xr¯utahe appears in Pt2, Ml4, P10, L4a,
L3, and P2 (sec. manu).
36
One further attestation of mr¯ura- may be seen in one of the fragments edited by W
ESTERGAARD
(FrW. 8,2: 1852-54, 334). The text as contained herein is much more parallel to the Pahlavi version
of Vd. 2,22 passage than this is itself; note the following equivalences: Vd. 2,22 haca. staxr¯o. mr¯ur¯o.
zii˚¯a: / FrW. 8,2 ha¯ec¯a. .. staxrahe. m e r e t¯o. zaiia.; Vd. 2,22 PT with gloss: stahmag¯ıh (¯ı-š [
〈ZY-š〉;
read
ziy¯a(-¯ı)?]
sturg k¯u harw gy¯ag andar šaw¯ed) / FrW. 8,2
auua\
a. staxr¯o. ya ˜t. h¯a. druxš. a¯eiti.; Vd.
2,22 PT with gloss: m¯udag (k¯u tis tab¯ah kun¯end) / FrW. 8,2 m¯oir¯os. ya ˜t. mahrk¯uš¯o. auua.m¯ırii¯aite.
> Vd. 2,22 PT with gloss: zamist¯an ras¯ed (*mark¯uš¯an g¯ow¯end) (> Npers. gloss in K1 [ed. B
ARR
/
I
BSCHER
1941, 34]
ism-i zamist¯an g¯oyand). Although the much distorted text of FrW. 8 deserves of
further investigation (according to W
ESTERGAARD
, it is found in two mss., viz. K15 and K38 [= M3
W
ESTERGAARD
]), it seems clear that the equivalent of
mr¯ur¯o is concealed both in
m e r e t¯o. (<>
mruta-?
cp. n. 34 above) and m¯oir¯os here. Should
〈m¯oir-〉 reflect 〈mr¯ur-〉 directly? — mruu¯aca in N.62 has
no context and cannot be taken into consideration (cf. H
UMBACH
1991, 89).
37
Similarly RV 1,51,15b; 5,3,12b; 8,40,12b; 10,54,6d; 7,58,6c may be added where traditional s´¯a
v¯aci conceals
s´¯a av¯aci. Cf. H
OFFMANN
1967, 219 sqq. for this restitution and for the function of the