Estudios Iranios y Turanios
Editores
Alberto Cantera
Juanjo Ferrer-Losilla
Número 5
Año 2023
Edita
SOCIEDAD DE ESTUDIOS IRANIOS Y TURANIOS
(SEIT)
Girona
Estudios Iranios y Turanios
Director:
Alberto Cantera
Secretario:
Juanjo Ferrer-Losilla
Comité de redacción:
Miguel Ángel Andrés-Toledo
Agustí Alemany-Vilamajó
Alberto Cantera
Juanjo Ferrer-Losilla
Götz König
Jaime Martínez-Porro
Éric Pirart
Depósito Legal: S-327-2020
ISSN: 2386-7833
Imprimida por: Printcolor
Ctra. de Mollet a Sabadell Km. 4,3 – Pol. Ind. Can Vinyals, Nave 18
08130 Santa Perpètua de Mogoda (Barcelona)
© Queda prohibida la reproducción total o parcial de los contenidos de este Boletín
sin permiso expreso de la Sociedad de Estudios Iranios y Turanios.
Envío de originales a:
Alberto Cantera, Fachbereich Geschichts- und Kulturwissenschaften, Institut für
Iranistik, Fabeckstr. 23-25, Raum 1.1024, 14195 Berlin (Deutschland). Correo-e:
alberto.cantera@fu-berlin.de.
Juanjo Ferrer, Département des Sciences de l’Antiquité, Langues et religions du
monde indo-iranien ancien, Bâtiment A1, Place du 20 Août 7, 4000 Liège (Belgique).
Correo-e: jjferrer@ulg.ac.be.
Índice de contenidos
Alberto CANTERA
The interpretatio iranica of Heterograms in Book Pahlavi:
The Case of YTYBWN- “To Sit Down, to Dwell and to Set”
and Some Related Problems ...................................................................... 7
Götz KÖNIG
Nicht-avestische Texte im Xorde Avesta:
die Texte des Danksagens ............................................................................ 23
Jaime MARTÍNEZ PORRO
7H[WDQG&RQWH[WRIWKH<DVQDƯ5DSLׇȕLQ ................................................ 47
Paolo OGNIBENE
About Some Kabardian Loanwords in Ossetic ............................................ 59
Éric PIRART
La vejez avéstica࣠ ...................................................................................... 65
The interpretatio iranica of Heterograms in Book Pahlavi:
The Case of YTYBWN- “To Sit Down, to Dwell and to Set”
and Some Related Problems
Alberto Cantera
Freie Universität Berlin
Chapters 18–23 on the verbal heterograms in the )UDKDQJƯ3DKODZƯJ࣠1 and the
OLVWRIYHUEDOKHWHURJUDPVLQWKH7XUIƗQIUDJPHQW2 provide only one Iranian equivalent for each heterogram. This is true even when in Book Pahlavi the same
heterogram is used for two homonymous verbs. In these cases, the Frahang provides only one Iranian verb as a meaning, the one that agrees with the meaning
of the heterogram: NTLWN- (cf. Aram. 1ܑ5 “to guard”) for SƗGDQSƗ\“to protect” and “to stand, to wait”; LHTWN- (cf. Aram. 5ۉ7 “to run”) for GDZƯGDQ
dawistan, daw- “to run” and “to speak” (daevic) (Nyberg/Utas 1988: 93). SimilarO\ZKHQWKHVDPHYHUEDSSHDUVZLWKGLIIHUHQWSDVWVWHPVHJâ5:1WQ'> for
IUƝVWƗGDQâ5:1\WQ'> for IUƝVWƯGDQ HJ:=â'5:1\W' IUƝVWƯG)
RU.7/:1WQ'> for PƗQGDQ.7/:1VWQ !IRUPƗQLVWDQ3 (e.g., PY4.11c, 23.3
.7/:1VW !PƗQLVW), the Frahang )Ư3 UHFRUGVDJDLQRQO\RQH —
the non-marked one, that is, the past stem with the phonetic complement -t(n)-.
This is also the surviving form in New Persian (PƗQGDQ IUHVWƗGDQ). The only
exception to this rule of the one-to-one relationship between heterogram and
LQWHUSUHWDWLRLUDQLFD is the well-known case of
. In Book Pahlavi, two different heterograms still distinguished in Epigraphical Middle Persian (EMP) fall
WRJHWKHU 2অ<':1 (03 2%<':1 kardan, kun- ³WR GR´ DQG 2+':1
(03 2+':1 JULIWDQ JƯU “to take”. Interestingly, the )UDKDQJ Ư 3DKODZƯJ
OLVWVWKHPWZLFH,Q)Ư3ZHILQGWKHKHWHURJUDPIRUJULIWDQJƯU2+':1WQ
JOSWQ 2+':1\W J\O\W 2+':1\P J\O\P 7KH VDPH VKDSH LV DOVR IRXQG
for kardan, kun-, however, strikingly, two different heterograms are listed here
)Ư3 2%<':1WQ %ে<:1WQ NUWQ 2%<':1\W %ে<:1\W NZQ\W
2%<':1\P%ে<:1\PNZQ\P$V1\EHUJQRWLFHVWKHVHFRQGRQHZKLFK
he interprets as Aramaic EQ\ “to build”, is never used in the known Iranian texts
(NYBERG/UTAS 1988: 49, 88).
This tendency within the Frahang has been continued in the modern standard
lexicographical works for Book Pahlavi. Thus, MACKENZIE (1971) and NYBERG
(1974) regularly indicate just one LQWHUSUHWDWLRLUDQLFD for each verbal heterogram.
1ௐ
On the Frahang s. (Heinrich F. J. JUNKER 1912; Heinrich Franz Josef JUNKER 1955; H. S. NY1988; MACKENZIE 2000; SHAKED 1993; KÖNIG 2022).
2ௐ
Fragment published by GELDNER and facsimiled by BARR (Karl Friedrich GELDNER 1904; BARR
1936).
3ௐ
Again in this case, the )UDKDQJƯ3DKODYL attests only one of the two possible LQWHUSUHWDWLRQHV
iranicae. ,WUHFRUGVRQO\ )Ư3 .7/:1WQ PގQWQ PƗQGDQDQG )Ư3â'5:1WQ
SO\VWގWQ IUƝVWƗGDQ/. The Frahang always choses, interestingly, the standard form in New
Persian.
BER/UTAS
Estudios Iranios y Turanios
Nevertheless, in this paper I will demonstrate that this one-to-one relationship between heterogram and LQWHUSUHWDWLRLUDQLFD is less universal than assumed. Some
heterograms can be read in different ways, despite the Frahang’s choice of one
reading. There are three main reasons for the possibility of multiple readings for
one heterogram. Firstly, sometimes the verb represented by the heterogram appears in different phonetic variations that might be either dialectal or chronological. Secondly, sometimes a heterogram covers different present stems of the same
root. They may or may not have a different value. The most famous example of
the former is the heterogram YTYBWN-, used for the intransitive verb nišastan,
QLãƯQ “to sit down,” as well as for the transitive verb QLãƗVWDQQLãƗQ “to set, to
establish”)4. Thirdly, different (partial) synonyms can be used for reading one verbal heterogram. Examples of this third exception are not recorded in the available
lexicographical works, but, as I will show in this paper, the heterogram YTYBWN“to sit (down), to dwell, to set” illustrates this possibility, since it may be read not
only as nišastan, QLãƯQ-, QLãƗVWDQQLãƗQ-, but also as ƗKLVWDQ, ƗK-.
Before I move on to the problems posed specifically by the heterogram
YTYBWN-, I will provide additional examples of the first two types of multiple
readings for a heterogram (different present stems and phonetic variants of the
same stem). The phenomenon of multiple present stems for one verb is well-known
in Old Iranian. In Middle Persian it is less frequent, but certainly not unknown.
Thus, in Manichaean Middle Persian, the verb ZLãƗGDQ “to untie” has two different present stems: ZLãƗK (ZLãƗKƝGZ\ãގK\G!FI3DUWKZLãƗK) (DURKIN-MEISTERERNST 2004: 359a; JOHNNY CHEUNG 2007: 136) and wiš- (ZLãƝG <wyyšyd>). In
Book Pahlavi the heterogram â/<7:1 is most often used. MACKENZIE reads the
present stem as *ZLãƗ\ led by NP JXãƗGDQJXãƗ\. Nevertheless, NYBERG (1974:
215) records both present stems for Book Pahlavi as well: ZLãƗ\ and wiš-. The
)Ư3UHFRUGVIRUWKHKHWHURJUDPâ/<7:1WKHSUHVHQWIRUPVZã\WZã\P for
ZLãƝGand ZLãƝP, but, according to NYBERG (1974: 215), the codices U2, U4 show
ZãҴ\W ZãҴ\P5, probably for ZLãƗ\ƝG ZLãƗ\ƝP. Both present stems are also atWHVWHGLQ3Ɨ]DQG,Q=:<ANTIA (1909: 344) has YƗãƗLWIRUâ/<7:1\W!
in the Pahlavi version, cf. ms. J3 YƗãƗƝW. These forms represent ZLãƗ\ƝGZLãƗҴƝG.
7KH)Ư3WUXHWRWKHLQWHQGHGRQHWRRQHFRUUHVSRQGHQFHUHFRUGVRQO\WKHHWHRJUDSKLFDO
spelling of the intransitive one: QLãDVWDQ QLãƯQƝG QLãƯQƝP Nonetheless, the double use of
the same heterogram already appears in Inscriptional Middle Persian, though with a difIHUHQWLDWLRQIRUWKHSDVWVWHPWKDWZHGRQRWILQGLQ%RRN3DKODYL<7<%:1VWIRUnišast
DQG <7<%:1W IRU QLãƗVW (GIGNOUX 1972: 1:38). Nyberg (and Utas) are of the opinion
that the same heterographic spelling could represent two originally different heterograms:
<7<%:1 IRU WKH LQWUDQVLWLYH QLãDVWDQ QLãƯQ would continue a participle passive \eWթƯEࡃ,
whereas the transitive form could be interpreted as \HܒD\\HEEթnjQLPSHUIHFWRI৫<%³WRPDNH
ready, to prepare” (NYBERG/UTAS 1988: 92). Dubious is the case of ških- “to be split, shattered”. B. PESCHEL RUDOO\ SRLQWVRXWWRPHWKHIRUP7%/:1\K\W!LQ<,WLVGXELRXV
as to whether it must be read as a new passive present stem ãNLQQLKƝG or rather as the original
and well-attested present stem ških- (<*šNDGL֒ D see below). In the latter, more likely case, the
heterogram TBLWN- was used for the transitive stem škenn- and the passive intransitive ških-.
It would follow the model of other verbal heterograms used for the passive with the phonetic
complement -ih-. Alternatively, we could read it as ãNLQQLKƝG. It would be a secondary passive
stem beside the old ških-.
5ௐ
However, these variants are not recorded in the edition (NYBERG/UTAS 1988: 21–22).
4ௐ
8
A. Cantera, The interpretatio iranica of Heterograms in Book Pahlavi
Nonetheless, in $\ƗGJDUƯ-ƗPƗVSƯJ, Annexe A, AGOSTINI edits the correspondence
RI â/<7:1\W! LQ WKH 3DKODYL YHUVLRQ DV YDVΩթW (proposing to correct it into
ZLãƗ\ƝG) (AGOSTINI 2013: 169). However, the ms. 5-19219 has the correct form
YDãΩթW, certainly for ZLãƝG࣠6. Thus, there is little doubt that even in the later phases of
transmission, Book Pahlavi knew both present stems. The present stem wiš- continues the old present X֒Lã+L֒ Dwith cognates in Avestan (Av. 3rd SSOVXEMKLLąQ
“they will bind” Yt8.55) and Vedic (YL«V\iWL). However, the stem ZLãƗ\ZLãƗK
( X֒LãD+L֒ D) does not have any correlates in the other Indo-European languages7.
The same verb also offers interesting examples of phonetic variants we might
HQFRXQWHU)LUVWO\WKH)UDKDQJWKH3Ɨ]DQGIRUPVDQG1HZ3HUVLDQVKRZWKHSUHVHQW
stem ZLãƗ\, but the MMP equivalent isZLãƗK. Both variants likely show different
ways to treat the glide for *wišƗҴƝPZLãƗҴƝd, HWF VHHWKH3Ɨ]DQGIRUPVYƗãƗLW֓࣠ ,
YƗãƗƝW). The tendency is for the disappearance of the \, but the morphological
boundary prevented the regular treatment. For preserving the hiatus, either the \
was preserved (against the expected evolution) or an h was introduced instead. This
phenomenon is also well-known in other Pahlavi words, although the exact rules for
its distribution (chronological or dialectal) remain obscure8:HILQGWKHYHU\VDPH
alternance in other verbs ending in either i֒ or d, etymological or going back to jғ.
Accordingly, we might find an alternation between h and \ or l į 910111213
i
uz-i
“to go out”
y (l)
Phl. X] L\ ƝG࣠9 ގZ]\W! := ,
X]L\ƝQGގZ]\G\QG! :=
MMP DGLKƝGގG\K\G!
ati-i֒- “to enter,
to approach”10
Ɨ-i “to come”
*þDL֒D“to mourn”
*u֒ LNƗL֒ am
“witness”
11
h
MMP X]LKƝGގZ]\K\G!
X]LKƝQGގZ]\K\QG!
Phl., MMP, NP Ɨ\- !\ ގ ގ
MMP þH\- <cy->
Phl. þHK- <cyh->
MMP JXJƗ\JZJ!\ގ
Psalter KXJƗ\ƯKJZNގG\K\!
3Ɨ]JXȖƗƯ, Arm. (loan) YND\,
Phl. JXJƗ\JZN!\ގ12
Phl. JXJƗKJZJގV!13,
Parth. ZLJƗKZ\JގK!
NP gugƗK
d
*ãNDGL֒D“to be broken”
MMP ških- <škyh->,
Judeo-Persian ãN\K-, škh-
6ௐ
Confusions between š and s are frequent in the Indian manuscripts and should not surprise us.
It could perhaps be explained as analogical to SƗGDQSƗ\ “to protect” (< SƗL֒ D) or ]ƗGDQ
]Ɨ\ “bear offspring, be born” (<*]ƗL֒ D) or even the pair QLãL\“to sit” and QLãƗ\ (see below).
8ௐ
Notice that most of the verbs with present stems in Ɨ\- do not show this alternance between
h and \ (e.g., IUDPƗ\SƗ\, etc.).
9ௐ
For this form, see below.
10ௐ
See OP DWL\ƗLã (JOHNY CHEUNg 2007: 155).
11ௐ
Kind indication by B. PESCHL.
12ௐ
In Phl. we find two different spellings: 1. JZN!\ގDQGJZNގV!7KHODWWHULVXVXDOO\
WUDQVOLWHUDWHGDVJZNގG\!DQGFRQVLGHUHGDSXUHO\JUDSKLFDOYDULDQWRIJZN!\ގ+RZHYHU
the existence of two different variants JXJƗ\ and JXJƗK is more likely.
13ௐ
See previous note.
7ௐ
9
Estudios Iranios y Turanios
*sid-i֒ a“to be destroyed”
Gଏғ
*hu֒ ƗGDL֒ a- <
*hu֒ Ɨ࠵̗ai֒ a-15
Dk abes-(L\);
NP gusil-14 < *wi-siį X֒i-sid-i֒ a-
Phl. abesih-ގSࡃVK!ގSࡃsh>, MMP
ގE\V\K!
[ZƗ\LãQKZގGãQ !16
Phl. [ZƗK17, MMP [ZƗK- [ZގK!
NP ࢷېƗK-
This evolution allows a satisfactory explanation for the passive suffix used for the
passive and denominative (CANTERA 2023). 14151617
The regular evolution of the group \ before endings with -Ɲ if the morphological boundary is not respected seems to have been Ɲ. Thus, the present wiš“untie” (see above) goes back to ZLãL\ƝG < *X֒Lã+L֒ D. Usually, the evolution
does not appear regularly through the whole flexion, but only in some forms of
the conjugation, mostly in the 3rd p.sg.: e.g., Phl. X]ƝG ގZ]\W! ³KH JRHV RXW´
< *X]L\ƝG (if this is not derived from the original athematic form *X]DL֒ WL); Phl.
QLãƝG “he sits” (see below) < *QLãL\ƝG.
Beside the variation between ZLãƗ\ZLãƗK, the witnesses of this verb show a
further phonetic variant concerning the initial consonant. In late Pahlavi, mainly
in the ritual instructions of liturgical manuscripts, we frequently find JXãƗGDQ
JZãގWQ NZãގWQ ! HJ 1HU< 9DU E JZãҴWQ in mss. 6, 15, 40, 2106, but
PV â/<7:1WQ EHVLGH ZãގWQ ! VHH ZãҴWQ ms. 5020 [year 1723] in
NerY59.29). The oldest known manuscript, ms. 2000 (K7) (1278, Rostam
0LKUƗEƗQ DOUHDG\VKRZVkwšҴWQ (NerVrS15.1 Var a). Nonetheless, this variant is
certainly quite modern and due to the influence of New Persian. The change of wito gu- is regular in Pahlavi only before m (see MMP, Phl. JZPҴQ “doubt”, MMP
JZP\[WQ, Phl. JZP\KWQ HWF ,QWKH3Ɨ]DQGKRZHYHUWKLVFKDQJHLVUHJXODULQ
all positions. The Pahlavi of the ritual instructions shows both possibilities. Accordingly, a dictionary of Pahlavi should include at least the infinitives ZLãƗGDQ
JXãƗGDQ and the present stems ZLãZLãƗ\(perhaps also ZLãƗK) and JXãƗ\, all of
them covered by the heterogram ŠLYTWN-.
Evidence of (partial) synonyms hidden behind a heterogram is far more difficult to find. The )UDKDQJƯ3DKODZƯJ, as we have already mentioned, listed only
one Iranian verb for each heterogram. Accordingly, the only chance of identifying
further alternative readings of manuscripts is the variation in spellings between
manuscripts. If in one passage one or more manuscripts show an eteographical
spelling different from the one recorded in the Frahang and the other manuscripts
show a heterogram, this indicates that the heterogram was not only read as listed
in the Frahang, but knew alternative readings as well. A systematic study will
only be possible when complete transcriptions of the most important Pahlavi manuscripts are available and searchable, a goal we pursue in the Middle Persian
14ௐ
It is used transitively (“to break off”), like the present stem wisinn-.
KX֒ƗGDL֒ D KX֒ƗMDL֒
ғ D-.
16ௐ
MACKENZIE reads iWDV[ZƗKLãQ.
17ௐ
,Q)Ư3WKHKHWHURJUDPBOYHWN- is read as KZҴVWQ KZҴK\W and KZҴK\P for [ZƗVWDQ
[ZƗKƝGand[ZƗKƝP.
15ௐ*
10
A. Cantera, The interpretatio iranica of Heterograms in Book Pahlavi
Corpus and Dictionary. Even then, it is unlikely that we will find many examples
of this type of variation. The Frahang’s rule of indicating only one LQWHUSUHWDWLR
iranica for each heterogram seems to be the consequence of a normative reading
of the Pahlavi that encompassed homogeneous readings for all heterograms, so
old alternative readings were annulled in favour of the normative ones.
Nonetheless, the example of the LQWHUSUHWDWLR LUDQLFD of the heterogram
YTYBWN- “to sit (down), to set” shows that such alternative readings existed and
that some of them had survived until the time of the manuscripts’ production. In the
following, I will analyse the LQWHUSUHWDWLRQHVLUDQLFDHpossible for this heterogram.
It displays some special features that make its analysis particularly attractive. Firstly, we have already seen that it might be used for the transitive and intransitive
present stems. Secondly, this is one of the verbs that has a different present stem
in Manichaean Middle Persian (QLãƯ\QLãƗQ) and New Persian (QLãƯQQLãƗQ).
$FFRUGLQJ WR )Ư3 DQG VRPH HWHRJUDSKLFDO VSHOOLQJV 3DKODYL ZRXOG PDWFK 1HZ
Persian and the present stem should be QLãƯQ. However, a systematic check of all
attestations has revealed spellings indicating the existence of an additional present
stem, QLãƯ\ beside QLãƯQ. Furthermore, a passage of the Pahlavi Yasna repeated
several times consistently shows an alternative reading, revealing that this heterogram could also be read as a different verb than nišastan orQLãƗVWDQ
Avestan uses two different verbal roots for expressing the notion of “to sit”,
both with well-established Indo-European etymologies:
1. had-: Pres.-stem: KLįD- < *KƯGD- < *si-sd-e-. It is frequently used with the
preverb ni-, and the resulting present stem is QLãKLįD-18
2. ƗK-: root present ƗK-: ƗVWH (3rd p.sg.), ƗծƾKƗLUH, ƗƾKΩ۬WH (3rd p.pl.), ƗKƯãƗ
(2nd p.sg. opt.), ƗծƾKƗQD- (part.pres.med.), stative: “to be sitting, to dwell”
(:$LD,
PIRART attempts to explain all forms of the present stem KLįD as belonging to the root hid- “to
reject”, Ved. sidh- (Pirart 1992: 2:80). The only attestation of the root had- would be Y10.15
QLãKDįDLWL and the present stem would be QLãKDįD and not QLãKLįD. Av. QLãKDįDLWL is, indeed,
the best attested reading, but the Sanskrit transmission has QLãKLįDLWL and °KLį is also attested
in other Indian manuscripts (QLãKLįΩWL: ms. 110; QLãKLįLWL: ms. 234 or the oldest Indian 9ƯGƝYGƗG manuscripts 4200, 4210, 4220, 4240). Furthermore, the equivalent of the Vedic present
stem VƯGiWL is the only form continued in Middle and New Iranian. MMP QLãƯ\- <nšyy>, Parth
QLãƯO <nšyl-> (JOHNY CHEUNG 2007: 125) indicates an original QLãƯGD. The quantity is, indeed,
confirmed by NP QHãƯQ (see below). Accordingly, the Avestan present stem KLįD shows a
shortening from KƯįD and we must reconstruct an I-Ir. present stem VƯGD despite the difficulties posed by this form. Notice, furthermore, that a similar alternation between QLãKLįD and
QLãKDįD is also found in other attestations that, according to PIRART, belong to the root hid:
18ௐ
nišKLįDƝWD
4000, 4050, 4161; 4200, 4210; 4610
4040, 4161; 4200, 4230, 4240
4040, 4161; 4200, 4210, 4240
4040, 4161; 4200, 4210, 4240
nišKLįƗW֓
V16.1 4000, 4010, 4025, 4040, 4050; 4600
V8.11
V9.33
V9.34
V9.35
nišKDįDƝWD
4010, 4025, 4040; 4230, 4240; 4600a
4000, 4010, 4025, 4050; 4210; 4600, 4610
4000, 4010, 4025, 4050; 4230; 4600, 4610
4000, 4010, 4025, 4050; 4230; 4600, 4610
niš.KDįƗW
4200, 4210, 4230, 4240; 4610
11
Estudios Iranios y Turanios
$FFRUGLQJWRWKHVWDQGDUGZRUNVLQ:HVWHUQ0LGGOH,UDQLDQRQO\WKHILUVWURRW
has survived, and there are no traces of the second. Forms of this root are found
only in Old Khotanese ƗK- (EMMERICK 1968: 20:13; JOHNY CHEUNG 2007: 154).
Furthermore, the present stem QLãƯGDwith preverb ni- is regularly continued in
Manichaean Middle Persian and Parthian. The heterogram YTYBWN- is used in
Book Pahlavi as intransitive as well as transitive, and the reader must choose how
to read it depending on his understanding of the passage. Inscriptional Middle
Persian still distinguishes between both, at least in the past stem, using the phonetic complement °st(n) for the intransitive stem and °t(n) for the transitive. The
present stems are usually read in Book Pahlavi as QLãƯQ andQLãƗQ (MACK ENZIE
1971: 60; NYBERG 1974: 142), exactly as in New Persian, whereas MMP shows
QLãƯ\- and QLãƗ\. The shape of the present stem postulated for Pahlavi is identical
to the standard form in New Persian: nišastan, nišin-. Such a stem is also attestHGLQRWKHUPRGHUQ:HVWHUQ,UDQLDQODQJXDJHVHJ*LO 5VKW ništœn, nišin-,
Shamerz. °ništ-am / näšniP, Sist. UãƯQ19, Siv. UãƯQQƯի ãW.
The origin of the present stem QLãƯQ is obscure. HÜBSCHMANN (1895: 101) derived NP nešin- from ni-šidQƗ. N. SIMS WILLIAMS (1981: 166) considers instead
a phonetical modification of the present stem QLãƯG D (see below) through an
assimilation of the nasal in distance. He compares it with other similar assimilations of the nasality in distance like those in the personal pronoun of the 1st
SVJ003an < **anam, Old Persian adam; Parth. QE\Q “cunning” beside GE\Q,
NP SDQƗP “face mask” beside Phl. SDGƗP (cf. Av. SDLWLGƗQD), etc., thus combining evidence from very different chronologies. Accordingly, he also explains
the transitive form QLãƗQ (present of QLãƗVWDQ) < *QLãƗGDL֒ D, Av. QLãƗįDLLDOP
QL\DãƗGD\DP, MMP QLãƗ\Qã!\ގ+RZHYHUWKHFKURQRORJLFDODQGGLDOHFWRORJical distribution between QLãƯQand QLãƯG (that is, QLãƯ\ or QLãƯį) and the date
postulates for the assimilation are not completely clear. The present stem QLãƯGD
is also preserved in most variants of Middle Persian and in Parthian. The stem
niãƯQ- DSSHDUV RQO\ LQ VRPH 1HZ :HVWHUQ ,UDQLDQ ODQJXDJHV DQG 3DKODYL 7KH
latter is the only Middle Iranian language attesting it. Yet, we cannot exclude the
possibility that in Pahlavi, this form is influenced by New Persian. If QLãƯQ is a
modern form, we could look for its origins only in Northwestern areas in which
the intervocalic d was preserved asį, but not in the dialects with the stem QLãƯ\.
In Pahlavi, we would expect QLãƯ\ (as attested in Manichaean Middle Persian) or
*QLãƯK(cf. the variants X]L\ƝQG and X]LKƝQG, see above).
)XUWKHUPRUHLQ%DORþLWKLVYHUEDSSHDUVDVQLãWQLVWQLQG whose present stem
seems to go back to a similar form, QLã Ư QG that could also perhaps explain
NP QLãƯQ and other forms (KORN 2005: 127). The relation between QLVWQLãWand
nind- seems comparable to that of sist, sind- “to break”, where the nasal stem is
inherited (Oind. FKLQiWWL, lat. scindo, etc.). In fact, it is not to be excluded that the
verbs °sistan, °sinn- and škastan, šken(n)- might have provided a model for the
creation of nišƯQ- beside QLãƯ\.
As we have seen above, the verb škastan knows two different present stems:
19ௐ
:LWKGLVVLPLODWLRQRIn..n to r..n (JOHNY CHEUNG 2007: 126).
12
A. Cantera, The interpretatio iranica of Heterograms in Book Pahlavi
šken(n)- (<*škand(a)-) and škeh- (< *škDGL֒ D), in a transitive/intransitive distribution. Similarly, we know that beside the present °sinn- (< *üLQG(a)-, cf.; Phl.
wisistan, wisinn-), a present °sih- (Phl. abesih-) and °sil- (<*üLGL֒ D-, Gr. ıȤȓȗȦ;
NP gusistan, gusil- “to shatter , to break”) (JOHNY CHEUNG I H[LVWV
that let us also suppose the existence of °VL\. The latter is originally intransitive/
passive, but it is used as transitive in New Persian (and also in Greek). In this
case, both present stems find cognates in other branches: *üind-a- (Lat. VFLQGǀ)
and VLGL֒ D(OInd. FKLG\iWH³WREHEURNHQ´*UıȤȓȗȦ 7KHVHSDLUV°sin(n)- VL\
sih- (abesih-) and škenn-, *ãNH\, might have served as a model for the creation of
QLãƯQ beside QLãL\ (/*QLãƯK) and for QLãƗQ beside QLãƗ\.20
škastan “to break, to be broken”
(škH\)škeh- “to be broken” škenn- “to break”
wisistan “to split”
*ZLVL\20(wisih-)
wisinn- “to split”
nišastan “to sit”
QLãƯ\(QLãƯK) “to sit”
>>> QLãƯQQ“to sit down”
QLãƗVWDQ “to set”
QLãƗ\/(QLãƗK) “to set”
>>> QLãƗQQ“to set”
Admittedly, this parallelism works well only at the formal level, but not at the semantic level. The “original” form alternating between °L\ and °inn- are in a transitive/
intransitive distribution (although ZLVL\ in NP gusil- is used as transitive). By
contrast, for QLãDVWDQQLãƗVWDQ, the \ and the –n- present are used as either transitive or intransitive depending on the root degree, and not on the final consonant of
the stem. Nonetheless, the inherited21 pattern of the intransitive °L֒ Dpresent beside
the transitive nasal present might have triggered the form QLãƗQ instead of the older
QLãƗ\ then the creation of QLãƯQ beside QLãƯ\. In any case, this analogical model
seems more likely than an unlikely change from d to n that should have occurred
before the change from d > \. The witnesses, however, are all modern, except for
Book Pahlavi, but here the influence of New Persian cannot be excluded.
Let us have a look at the distribution of the forms in the variants of Middle
Persian. The forms attested in MMP are exclusively QLã\\ and QLãҴ\. This is also
the present stem attested in the Middle Persian of the Psalter, where, beside the expected ideographical spelling YTYBWN-, at least twice we find the eteographical
spelling QLãƯ\22:
QLãƯ\ƗQ Qã\GގQ! ³, VKDOO VLW´ 3V QLãƯ\ƗQ DEDULW JƗKǀJ “I shall sit
upon your throne”
QLãƯ\ƗQGQã\GގQG\!³WKH\VKDOOVLW´ 3V ƝJLãƗQIUD]DQGL]QLãƯ\ƗQG
SDVWǀDEDUJƗKǀJƯWǀЋƗ\ƝGƗQƯЋƗ\ƝGƗQ “their children shall also sit after
you upon your throne forever and ever”
GIGNOUX also assumes this reading for the heterogram in the Middle Persian inVFULSWLRQV<7<%:1G!QLãƯ\ƝQG ..=.15P GIGNOUX
1972: 1:38). The reading could, of course, also be QLãƯQƝQG, if we rely on the
20ௐ
See NP °sil- “to split” < °VLį. This form is originally intransitive, but also used as transitive.
The Vedic verbs in \i also produce secondary nasal stems such as ĞLQiܒ܈L, ĞLۨ܈DWL (Br +)
from ĞL\܈i- (KULIKOV 2011: 195).
22ௐ
:HDOVRILQGWKHIROORZLQJKHWHURJUDSKLFDOVSHOOLQJ<7<%:1W!nišast (99.1).
21ௐ
13
Estudios Iranios y Turanios
comparison with the supposed Pahlavi QLãƯQ rather than with MMP and Psalter nišƯ\. Pahlavi would agree with New Persian and disagree with Manichaean
Middle Persian and Psalter:
Parth
QLãƯį<nšyd->
MMP
QLãƯ\- <nšyy->
QLãƗ\QLã!\ގ
Psalter
Phl
NP
QLãƯ\- <nšyd-/ QLãƯQ <nšyn- / QLãƯQ
<7<%:1!
<7<%:1!
23
QLãƗOQãގO!
QLãƗQQ\ãގQ QLãƗQ
<7<%:1!
The evidence in Book Pahlavi for reading the present stem represented by
YTYBWN- as QLãƯQ and QLãƗQ and not as QLãƯ\ and QLãƗ\ is from a late dating: the )UDKDQJƯ3DKODZƯJVFDUFHHWHRJUDSKLFDOVSHOOLQJV3Ɨ]DQGDQG3ƗUVƯJ
The greatest weight is likely attributed to the )UDKDQJ Ư 3DKODZƯJ )
Ư3
<7<%:1WQ nšstn<7<%:1\WQã\Q\W<7<%:1\PQã\Q\P. This reading
seems to be confirmed by the very rare cases in which some manuscripts shows
the eteographical spelling. I have been able to find following further rare eteographic spellings among the overwhelming majority of heterographical ones:23
Qã\Q\W /QLãƯQƝG39 PVV *U%G 4ࡃ'0Qã\Q\W7'7'24)
Qã\QP /QLãƯQDP/ GrBd4.8 (mss. DH, TD1, TD2)
For the transitive form, we also found some eteographical spellings, such as:
QãҴQ\W /QLãƗQƝG/ Y9.24 (400, 500, 510)
QãҴQQG /QLãƗQƝQG*U%G 7'QãގQQG7'QãގQG
)XUWKHUPRUHWKLVLVWKHVWDQGDUGUHDGLQJZHILQGLQ3Ɨ]DQGDQG3ƗUVƯJ$IHZ
examples will suffice:
3ƗUVƯJQLãƯQƝG $Ư- QLãƯQDQG $Ư-
3Ɨ]DQG =:< QLãƯQΩ۬W (Antia 347), QLãnjQΩ۬G (ms. J3, p. 112), QLãƯQΩթW
QLãƯQDQGD7KH3ƗUVƯJYHUVLRQRIWKH$\ƗGJDUƯ-ƗPƗVSƯJ shows the forms of
QLãƗVWDQ, QLãƯQDOUHDG\LQWKHLU1HZ3HUVLDQIRUP:HILQGnišast, nišastand,
QLãDVWƝG, QLãƯQDP, QLãƯQƝGQLãƯQDG (AGOSTINI 2013: 223).
However, we cannot exclude the possibility that this evidence just reflects the
influence of New Persian at the time of the manuscripts’ production. A few deviating spellings indicate the existence of the present stem QLãƯ\- beside QLãƯQ. In
four passages (double as many as those containing the eteographical spellings of
QLãƯQ), we find spellings that suggest a reading as QLãƯ\, but as we will see, two
may perhaps be explained another way. In GrBd9.21, it is affirmed that the fire
%XU]ƝQPLKUVLWV <7<%:1\W! RQWKHPRXQWDLQ5ƝZDQG7KHYHUELVXVHGDV
the intransitive form of QLãƗVWDQQLãƗQ“to enthrone” a fire:
Ɨ QNXVWDJUƝZDQGNǀISDG[ZDUƗVƗQNƝãƗGXUƯEXU]ƝQPLKUSDGLãQLãƯQƝGQLãƯ\ƝG
³7KHPRXQWDLQ5ƝZDQGLQ;ZDUƗVƗQRQZKLFKVLWVWKHILUH%XU]ƝQPLKUEHORQJ
to this region.”
23ௐ
24ௐ
Only in Ps131.7 QLãƗOLãQ, cf. NP gusil- but3KOZLVL\-.
,QWKHVDPHOLQH<7<%:1GLVXVHGIRUQLãƗQƝQG (TD2 65v).
14
A. Cantera, The interpretatio iranica of Heterograms in Book Pahlavi
The manuscripts of the Iranian Great Bundahišn have the heterogram
<7<%:1\W!%\FRQWUDVWWKHPDQXVFULSWVRIWKH,QGLDQYHUVLRQKDYHDQHWHRgraphical form: (K20
; M51
. This is the standard spelling of the
past form nišast. But why should the Indian Bundahišn contain the verb in the
past that appears in the present in the Iranian Bundahišn? It is more likely that this
spelling represents also a present.
In the Pahlavi commentary to PV7.1225 we may also find an instance of a similar spelling. The passage runs as follows:
ZDVWDUDJ DNƗU ƝQ … ƗQ NƝ ]DQ Ư GDãWƗQ SDG PHKPƗQƯK GƗUƝG Xã DQGDU EƝ
QLãƯ\ƝGD\ƗEDãQDVƗDEDUUDVƝG
The following clothes are useless … the ones which a menstruating woman has
with her in isolation and sits on them or the ones a corpse reaches26.
MOAZZAMI reads the verb as “QLãƯQƝG QLKƝG” and offers no comment or expla, 4610
nation. The manuscripts show the following spellings: ms. 4000
and its 4615
, 4700 (copy of ms. 4600)
. None of these
attested spellings can be read as QLKƝG.
A similar form is found twice (PY1.3, 2.3) in ms. 400. There, we read QLãƝQG
Q\ã\QG!ZKHUHRWKHUPDQXVFULSWVVKRZ<7<%:1G!,Q3<WKHIRUPIROORZV<7<%:1G!DQGIXQFWLRQVDVLWVJORVVLQGLFDWLQJLWVLQWHUSUHWDWLRLUDQLFD
These two passages deserve a deep analysis, which I will offer later, but for the
PRPHQWLWVKRXOGVXIILFHWRQRWHWKDWWKHFRS\LVWRIPVUHDG<7<%:1G!
as Q\ã\QGThis can hardly be read differently than QLãƝQG(and not as QLãƯQƝQG,
since this scribe is completely systematic in the distribution of the ending of the
3UGSSO\QG in eteograms and -d in heterograms).
The form in GrBd9.21 could be interpreted as QLãƯ\ƝGQã\\W!:HZRXOGH[pect °<ydyt> (cf. Phl. JUL\ƝG <glydyt> < *JUֈGD L֒ D WL), but the graphical reduction
is comparable to spellings such as ãƗ\ƝGã\ގW!DEƗ\ƝGގSࡃ\ގW !%\FRQWUDVW
25ௐ
MOAZZAMI numbers this passage V7.13 because she adds after VS7.2 a Pahlavi translation
missing in the manuscripts. According to her, the Avestan text would correspond to VS8.16
and consequently, she adds the Pahlavi translation that she finds there. The liturgical manuscripts do, indeed, complete a text that does not appear in the Pahlavi translation. However,
this text is not identical to VS8.16, but the following text: YƯVSΩPƗDѩƗW֓VSƗYƗDLȕLYDƝQDLWL IUƗ YƗ [ࢷDUDLWL YDLLǀ YƗ KƯ NΩUΩIã[ࢷDUǀ DǀXL X]XXD]DLWHƗDW֓ DƝãD GUX[ã \Ɨ QDVXã XSD
GXXąVDLWL DSƗ[ΩįUDƝLELLǀ QDƝPDƝLELLǀ PD[ãL NΩKUSD ΩUΩȖDLWLLD IUDãQDǀã DSD]DįDƾKǀ
DNDUDQΩPGULȕLLƗծ. \DׇD]ǀLåGLãWƗLã[UDIVWUƗLã This passage is unknown elsewhere and was
omitted from the Pahlavi translation due to a transmission error: a typical saut de même
u même*HOGQHUDQQRXQFHVLQKLVHGLWLRQWKDWWKLV DQGRWKHUWH[WV DYDLODEOHLQWKH6ƗGH
manuscripts, but missing in the Pahlavi translation, will be edited in an appendix: “For the
additional passage of the Vd. sâdes, see the Appendix” (GELDNER 1886: 3.48). WESTERGAARD
includes the passage, but in brackets, and indicates that it is omitted in K1, 2, Bvd. (WESTERGAARD 1852: 387).
26ௐ
MOAZZAMI translates this passage as follows: “That with which a menstruating woman has
contact (= in her use), through her sitting on it.” She fails to translate SDGPHKPƗQƯK. This
expression is frequently used in the Pahlavi version of 9ƯGƝYGƗG for the place of isolation,
far from fire and water, e.g., PV5.56G DQGDUVƝšabagNƗUSDGPHKPƗQƯKQƝSƗGL[ãƗNDUGDQ
“For three days she (the menstruating woman) is not allowed to do any work in isolation”.
The same usage is also found in ŠnŠ2.96 ƗQNƝ]DQƯGDãWƗQSDGPHKPƗQƯKGƗUƝGSDGKDPNƗU
be hilišn “The clothes that a menstruating woman has with her in isolation should be left
exclusively for this use.”
15
Estudios Iranios y Turanios
PV7.12 could be read as QLãƝG and PY1.3 and 2.3 as QLãƝQG࣠27. They apparently
show the same reduction that we also find in the present stem wiš- of ZLãƗGDQ
)Ư3Zã\WZã\P for ZLãƝGand ZLãƝP and MMP ZLãƝG <wyyšyd>; see above).
These forms go back °X֒Lã+L֒ D so we would expect *ZLãL\ƝGƝP. Comparable is
the 3UGSVJX]ƝGގZ]\W!IRU X]L\ƝG FI003ގZ]\K\G!ގZ]\K\QG!X]ƯKƝG
X]ƯKƝQG (the same form, but with a different glide; see above). It is unclear whether this is just a graphical reduction for QLãƯ\ƝG X]L\ƝG (cf. ]D\ƝG “he asks for”
]\W !39)Ư2 RUDUHDORQH nƯãƝGX]ƝG), but MMP ZLãƝG <wyyšyd>
seems to indicate a real reduction.
PY1.3 and 2.3 are just two of many instances in a frequently-attested comPHQWDU\RQ0LࢡUD¶VHSLWKHWVZKHQKHLVPHQWLRQHGDVWKHSDWURQRIKƗXXDQL, the
PRUQLQJ RIWKHGD\ LQWKH=RURDVWULDQOLWXUJLFDOFDOHQGDU7KLVVHFWLRQLVUHFLWHG
in the Yasna in all mentions of the fiveJƗKs, that is, at least one time in each litany:
Y1.3, 2.3, 4.8, 7.6, 22.6, 24.13 and 66.3. In many cases, the text is not abbreviated,
and the Pahlavi translation and gloss are written in extenso in some manuscripts.
,QWKH3DKODYLWUDQVODWLRQDJORVVH[SODLQVWKHHSLWKHWVRI0LࢡUDKD]DƾUǀJDRãD
“with a thousand ears” and EDƝXXDUΩFDãPDQ-, “with ten thousand eyes”. ApparHQWO\WKH\GLVWXUEHGH[HJHWHVZKRNQHZUHSUHVHQWDWLRQVRI0LࢡUDWKDWGLGQRWILW
the Avestan description. Consequently, they emphasize that this is not a physical
IHDWXUHRI0LࢡUDEXWPXVWEHXQGHUVWRRGPHWDSKRULFDOO\7KH³HDUV´ZRXOG
simply mean that 500 PƝQǀ\³VSLULWV´VLWLQ0LࢡUD¶VKHDGIRUWKHIXQFWLRQRIKHDUing and, accordingly, the “10,000 eyes” belong to the 5,000 spirits that are there
for seeing. The text runs as follows (the verb “to sit, to dwell in” is left unread for
the moment, and appears as XXX):
$3ãJZã\K+1$$<ã'ࡃP\QZ\ގZ' LOYŠE XXX$3ãNގO<JZã\K
2অ<':1[$<1(QJZKã:=.QJZKã$3ãE\ZOFãP\K+1$$<ã
P\QZ\ގZ' LOYŠE XXX$3ãNގO<FãP\K2অ<':1[$<1(
+<7:1:=.+<7:1
XãKD]ƗUJǀãƯKƝGNnjãSDQЋVDGPƝQǀ\ǀVDU;;;XãNƗUƯJǀãƯKNXQƝQGNnjƝQ
QL\ǀ[ãXGƗQQL\ǀ[ãXãEƝZDUþDãPƯKƝGNnjãSDQЋKD]ƗUPƝQǀ\ǀVDU;;;Xã
NƗUƯþDãPƯKNXQƝQG
KD]ƗUJǀãƯK (means) the following: five-hundred PƝQǀ\ dwell in his head and
do the function proper of the ear: “hear this, hear that”. And EƝZDUþDãPƯK
(means) that five-thousand PƝQǀ\ dwell in his head and do the function proper
of the eye: “See this, see that”
As for the verb “to be sitting, to dwell”, two different spellings alternate in the
PDQXVFULSWV%HVLGHWKHKHWHURJUDP<7<%:1G!ZHDOVRILQGDQHWHRJUDSKLFDOVSHOOLQJގK\QG!
The presence or absence of \ in the preverb is less problematic than it may appear. The
preverb can, indeed, be written with (nidom “smallest” <nytwm>, QLKƗQ³VHFUHF\´Q\KގQ !
etc.) or without \(nibištanQSãWQ !³WRZULWH´nifridag³DFFXUVHG´QSO\WN !nihuftan “hide”
QKZSWQ !QLãƝP “seat” <nšdm>, etc.).
27ௐ
16
A. Cantera, The interpretatio iranica of Heterograms in Book Pahlavi
1st attestation
Y1.3
Y2.3
Y4.8
<7<%:1G
ގK\QG
<7<%:1G
ގK\QG
400: <7<%:1G
<7<%:1GQ\ã\QQG28
415: n\ãގQ G
ގKQ VHFP<7<%:1G
<7<%:1G
ގK\QG
<7<%:1G
ގK\QG
<7<%:1G
ގK\QG
<7<%:1G
ގK\QG
<7<%:1G
<7<%:1G
<7<%:1G
510’ގK\QG
<7<%:1G
<7<%:1G
<7<%:1\W
<7<%:1G
Y7.629 500: Y[...]B[.]Nd
Y22.6
2nd attestation
<7<%:1G
>«@7>@:1G
ގK\QG
Y24.13 <7<%:1G
<7<%:1G
<7<%:1G
<7<%:1G
Y66.3
The distributional pattern corresponds to the standard behaviour of heterograms
and eteograms:2829
a. in the same passage, the same manuscript has the heterogram once and the
eteogram once (500 in Y1.3, 2.3 and ms. 400 in Y22.6 and 66.13)
b. in the same passage, a manuscript has the heterogram while another one or
other ones has/have the eteogram (e.g., the 2nd attestation of Y4.8 in ms. 500).
,QWHUHVWLQJO\0LKUƗEƗQWKRXJKOLNHO\FRS\LQJWZLFHIURPWKHVDPHRULJLQDODOVR
alternates between both spellings: in the second attestation of Y4.8, he writes
<7<%:1G!LQWKHPVDQGODWHUގK\QG!LQ$SSDUHQWO\LWZDVVWLOO
FOHDUIRU0LKUDEƗQWKDWގK\QG!LVWKHHWHRJUDSKLFVSHOOLQJRI<7<%:1G!
7KHVFULEHRIPVVRPHWLPHVDWWHPSWHGWRDVVLPLODWHWKHVSHOOLQJގK\QG!
with the spelling of QLãƯQƝQG, as he clearly does in Y1.3, where he added this gloss
after the heterogram. For that purpose, the final stroke of <LOYŠE> is taken as the
initial n and -ҴK- is understood as \ã. However, for an eteogram, the ending of the
3rd SVJLVLQPV DVLQPRVWPDQXVFULSWV DOZD\V\QG, whereas the ending -d
is reserved for the heterograms. Thus, the reading in ms. 400 could just be read as
QLãƝQG. Accordingly, ҴK\QG could be understood as a misreading of QLãƝQG, 3rdSVJ
of a present stem niš- (see above). However, the manuscripts (except ms. 400)
28ௐ
29ௐ
Also 420, 450, 456, 457.
Abbreviated in most manuscripts.
17
Estudios Iranios y Turanios
FOHDUO\ZULWHގK\QG!DQGQRWQ\ã\QG!ZKLFKLVDOVRWKHlectio difficilior. The
attested spelling is most likely to be read as ƗKƝQG, 3rd SSORIWKHYHUEƗK-.
The verb ƗK- “to sit” is well-attested in Avestan (see above), but apparently it survives in Middle Iranian only in Old Khotanese ƗK- “to sit, to remain”
(EMMERICK 1968: 20:13; JOHNY CHEUNG 2007: 154) and in Bactrian ĮȡȝĮĮȣ“to
be present, to stay” (SIMS-WILLIAMSDJOHNY CHEUNG 2007: 154). This
gloss would also attest it in Pahlavi. Its presence in the gloss is indeed well-justified: the main semantic difference between the present QLãƯGD and ƗK- seems
to be that the first one is telic and alludes to the action of “sitting down”, whereas
the root present stem ƗK- is stative and means “to be seated, to dwell”. Although
in Avestan the present stem QLãKLįDmight already be used with a stative value,
the verb ƗK- inherently posesses the stative value and can never be used as telic
(“to sit down”), see N19.6 *DLLǀYƗWDFDYƗ KLãWΩPQǀYƗƗծƾKƗQǀYƗ SDׇƗQǀYƗ
EDUΩPQǀYƗYD]ΩPQǀYƗDLȕLLƗVWǀDׇDUDWXIULã “whether walking, running, standing, sitting, lying down, riding, or driving, as long as he wears the girdle, he
satisfies the ratus”; Yt10.45 \Hƾ̗KH DãWD UƗWDLLǀ YƯVSƗKX SDLWL EDUΩ]ƗKX YƯVSƗKX
YDƝįDLLDQƗKX VSDVǀ ƗծƾKƗLUH “whose eight assistants stay on all hillocks, on all
watchtowers as spies”. In our passage, the verb is clearly stative and indicates
WKDWRUVSLULWV³DUHGZHOOLQJ´LQ0LࢡUD¶VKHDG:KHWKHUWKH\DUHVHDWing or standing is certainly irrelevant, and the passage certainly does not indicate
that the spirits are now sitting down. Hence, we might conclude that behind the
heterogram YTYBWN- is hidden not only the Iranian intransitive verb QLãƯQ “to
sit down” and the transitive QLãƗQ “to set down, to establish” (frequently used
regarding fire), but also the stative verb ƗK-.
The past stem ƗKLVW is also attested in the zand, as already acknowledged by
BARTHOLOMAE ,Q9=DUDࢡXãWUDLVSUHVHQWHGDVDVNLQJDTXHVWLRQ
ZKLOHVLWWLQJQH[WWRWKHPHDQGHURIWKH'DUԥۥƯ SΩUΩVDW ]DUDׇXãWUǀGDUϷMLLDSDLWL
]EDUDKHDKXUƗLPD]GƗLYDƾKDXXHYRKXPDLWHƗծƾKDQǀ). The zand interprets DKXUƗL
PD]GƗLYDƾKDXXHYRKXPDLWHas the subject of the verb “to sit”:
SDGGUƗML]EƗUNnjǀKUPD]GXGZDKPDQƯZHKƯZHKSD\PƗQƗKLVW KƝQG ašawahišt
[ãDKUHZDUXGVSDQGDUPDG
2QWKHPHDQGHURIWKH'UƗMLZKHUH2KUPD]GDQGWKHJRRG:DKPDQRIJRRG
PHDVXUH$ãDZDKLãW;ãDKUHZDUDQG6SDQGDUPDGZHUHVLWWLQJ
The past participle is written in Avestan characters in mss. 4600
(
) and 4610 (
).
Despite the scarcity of eteographical spellings, the existence of the Pahlavi
verb ƗKLVWDQ, ƗK- is certain, as it is also certain that the heterogram YTYBWN- was
used not only for QLãDVWDQQLãƯQ, but also for ƗKLVWDQ, ƗK-. Thus, it is likely that
some of the attestations of YTYBWN- that we traditionally interpret as witnesses
of QLãDVWDQQLãƯQ were indeed read as ƗKLVWDQ, ƗK-, at least at some point. Unfortunately, it is impossible to decide upon the distribution. In all eteographic attestations, ƗKLVWDQ, ƗK- clearly has a stative value “to be seated, to dwell”. However,
QRRWKHU:HVWHUQ0LGGOH,UDQLDQODQJXDJHNQRZVƗK-, and these others use QLãƯ\
QLãƯį for the stative as well. Thus, a widespread use of ƗK- in Pahlavi is unlikely.
18
A. Cantera, The interpretatio iranica of Heterograms in Book Pahlavi
Rather, the attestations seem to be an archaism of the zand.
An archaism is, indeed, expected in the zand of the 9ƯGƝYGƗG, but not so much
in the zand of the litanies of the Yasna, since this is the part of the Yasna that was
OLNHO\QRWLQFOXGHGLQWKH6WǀG<DVQDQGZKRVHWUDQVODWLRQZDVSHUKDSVUHGDFWHG
at a later stage. Nonetheless, it is very likely that there was also an old zand of the
dedications, since they contain the official names and epithets of the gods. In any
case, the knowledge of the zand RI0LࢡUD¶VGHGLFDWLRQZDVZLGHVSUHDG:HDOVR
find it in the Great Bundahišn (GrBd26.71–3), although unfortunately, the verb
“to dwell” is not used there:
þL\ǀQJǀZƝGNnjPLKUƯIUƗ[Jǀ\ǀGƯƝZKD]ƗUJǀãƯEƝZDUþDãPXãIUƗ[Jǀ\ǀGƯK
ƝGNnjNDSDGDãWDEƝEƯPƯKƗEƝãƗ\ƝG madan ud šudanSDGUƗKƯPLKUXãKD]ƗU
JǀãƯKƝGNnjãSDQЋVDGPƝQǀ\NƗUƯJǀãƯKKDPƝNXQƝGXãEƝZDUþDãPƯKƝQNnjã
SDQЋKD]ƗUPƝQǀ\NƗUƯþDãPƯKNXQƝQGNnjPLKUƝQL]QL\ǀãXGƗQL]QL\ǀãƝQL]
ZƝQXGƗQL]ZƝQ
As it says: “Mihr of wide pastures who has one thousand ears and ten thousand
eyes.” ‘Having wide pastures’ (IUƗ[Jǀ\ǀG) means that one can go without fear
in the dessert on the path of Mihr. ‘Having a thousand ears’ (KD]ƗUJǀãƯK) means
that five-hundred spirits play for him the role of the ears. ‘Having ten thousand
eyes’ (EƝZDUþDãPƯK) means that five thousand spirits play for him the role of
the eyes.
This analysis of the different possible LQWHUSUHWDWLRQHVLUDQLFDH of the heterogram
YTYBWN- shows that, against the impression transmitted by the lexicographical
works and even by the )UDKDQJƯ3DKODZƯJ, one heterogram could be read in different ways. These different readings could be simply phonetic variants (dialectal
or chronological), different stems of the same verb or even different verbs. The
one-to-one relation between heterogram and eteogram in the Frahang is artificial, didactic and normative. The selection of the verbal form was likely even
influenced by New Persian. It seems that New Persian had notably influenced
the way to read Pahlavi by the time of the manuscripts’ production and they (as
ZHOODV3Ɨ]DQGDQG3ƗUVƯJ UHYHDOLWLQWKHLURFFDVLRQDOQRQKLVWRULFDOVSHOOLQJV
NYBERG’s transcription system (following BARTHOLOMAE’s proposal) simply relied upon the Pahlavi historical orthography and did not attempt to modernize it.
HENNING’s claim of liberating the transcriptions from the burden of the historical
orthography (HENNING 1958: 122) was then followed by MACK ENZIE’s successful
transcription system. He pretended to transcribe “the Middle Persian koine of the
HDUO\ 6DVDQLDQ HPSLUH WKH SHULRG RI .DUGƝU DQG WKH =RURDVWULDQ 5HIRUPDWLRQ´
(MACK ENZIE 1967: 18). The documents in Manichaean Middle Persian and Parthian provide us with the necessary tools for reconstructing this état de langue.
This assumes, firstly, that the Pahlavi language did not fundamentally change
from this date until the moment of production of most Pahlavi texts in the 9th
century — a rather unlikely assumption. Manichaean Middle Persian would be
(as already postulated by TEDESCOI MXVWWKH³SXUHU´FRXUWODQJXDJH
whereas Pahlavi as a “national” language displays more influences from other
dialects (such as Parthian), as also occurs with New Persian. However, the way
19
Estudios Iranios y Turanios
to read Pahlavi during the time of the manuscripts’ production (and also of the
3Ɨ]DQGDQG3ƗUVƯJYHUVLRQV ZDVOLNHO\LQIOXHQFHGE\1HZ3HUVLDQVRWKDWWKH
New Persian reading might have substituted for other older Pahlavi forms in the
manuscripts as well as in other tools at our disposal for the transcription of Pahlavi30. The historical orthography of Pahlavi creates a gap between the script and
phonetic realization that facilitates the modernization of the pronunciation without affecting the way to write, or at the very least, only slightly and occasionally.
This is certainly true for the reading of the heterograms, but it also affects the
eteographical spellings.
The same also applies for the )UDKDQJƯ3DKODZƯJ. The latter certainly continues a long-lasting tradition, but it was changed and adapted until a very late date. Firstly, the Frahang
Ư 3DKODZƯJ FRQWDLQV KHWHURJUDPV WKDW DUH QR ORQJHU XVHG LQ %RRN 3DKODYL OLNH â:0$/+
(aram. ãXQQƗUƗ) for Phl. gurbag “cat” (NYBERG/UTAS 1988: 73; MACKENZIE 2000: 125). SecRQGO\WKH7XUIƗQIUDJPHQWKDVDUHVSHFWDEOHDQWLTXLW\DVLWLQFOXGHVVXEMXQFWLYHIRUPVWKDW
were no longer in use in the classical Pahlavi of the 9th century (KÖNIG 2022: 98). Thirdly, the
JORVVDU\PHQWLRQHGE\,EQ0XTDIIDޏDVFRQWDLQLQJDWKRXVDQGKHWHURJUDPVZLWKWKHLU,UDQLDQ
equivalents is most likely a composition similar to our )UDKDQJƯ3DKODZƯJ(JAMASPJI-ASA/
HAUG 1870: 37; KÖNIG 2022: 194). Nonetheless, the )UDKDQJƯ3DKODZƯJ includes pseudo-heterograms “coined by later scribes from Arabic words” (like ASD-l for ãƝU “lion”, cf- Ar. asad)
(MACKENZIE 2000: 125) and New Persian words such as the late (New Persian) loanword
DPãƗVIDQG (Av. DPΩ܈հDVSΩ۬WD) instead of Phl. (a)PKUDVSDQG (KÖNIG 2022: 188). In fact, we
have to understand the )UDKDQJƯ3DKODZƯJ, as noted by SHAKED, as “an open-ended notebook” with a fluid transmission (SHAKED 1993: 75). Since the extant manuscripts used in all
editions go back to the 17th or 18th century (NYBERG/UTAS;,±;9 ZHFDQQRWH[FOXGH
the possibility that modernization of the readings has occurred and again, and of influences from New Persian, as happens in the famous case of the demonstrative pronoun (h)ƗQ:
)Ư3UHDGVƗQ (cf. NP ƗQ), but MMP KƗQKގQ!
30ௐ
20
E. PIRART Pour de nouveaux fragments avestiquesODJpQpDORJLHGH=DUDׇXãWUD
Bibliography
Agostini, D. (2013). $\ƗGJƗUƯ-ƗPƗVSƯJXQWH[WHHVFKDWRORJLTXH]RURDVWULHQ (Biblica et
orientalia 50). Roma.
$QWLD(. Pâzend texts. Bombay.
%DUU. ³5HPDUNVRQWKH3DKODYL/LJDWXUHV;DQG;´LQBulletin of the School
of Oriental and African Studies 8: 391–403.
Bartholomae, Chr. (1904). Altiranisches Wörterbuch. Straßburg.
&DQWHUD$ ³7KH3DVVLYH6XIIL[Ʈࡅ KLQ0LGGOH3HUVLDQ´LQIranica 33.
Cheung, J. (2007). (W\PRORJLFDO'LFWLRQDU\RIWKH,UDQLDQ9HUE (Leiden Indo-European
Etymological Dictionary Series 2). Leiden-Boston.
Cheung, J. (2007). (W\PRORJLFDO'LFWLRQDU\RIWKH,UDQLDQ9HUE(Leiden Indo-European
Etymological Dictionary Series). Leiden.
Durkin-Meisterernst, D. (2004). 'LFWLRQDU\RI0DQLFKHDQ0LGGOH3HUVLDQDQG3DUWKLDQ
'LFWLRQDU\RI0DQLFKDHDQ7H[WVVol. III. Texts from Central Asia and China,
Part 1). Turnhout.
Emmerick, R. (1968). Saka Grammatical Studies. Vol. 20 (London Oriental Series). London.
*HOGQHU.) Avesta. The Sacred Books of the Parsis. Stuttgart.
Geldner, .) ³%UXFKVWFN (LQHV 3DDKODYL*ORVVDUV $XV 7XUIƗQ &KLQHVFKLFK
7XUNHVWƗQ´LQSitzungsberichte Der Königlich Preussischen Akademie Der Wissenschaften: 1136–37.
Gignoux, Ph. (1972). *ORVVDLUH 'HV ,QVFULSWLRQV 3HKOHYLHV HW 3DUWKHV. Vol. 1 (Corpus
Inscriptionum Iranicarum. Suplementary Series). London.
+HQQLQJ :% “Mitteliranisch”, in: Handbuch Der Orientalistik IV. Iranistik.
Linguistik: 20–130.
Hübschmann, H. (1895). Persische Studien. Straßburg.
-XQNHU+)-(1912). The Farhang-i Pahlavik. 1 vols. Prolegomena. Heidelberg.
-XQNHU+)- 'DV)UDKDQJL3DKODYƯNLQ]HLFKHQJHPlHU$QRUGQXQJ:LHVEDGHQ
.|QLJ * ³)URP :ULWWHQ WR 2UDO" 7KH (QFRGHG 3DKODYL LQ WKH )UDKDQJ Ư
3DKODZƯJ´LQ$6KD\HVWH'RXVW HG 'ƗGHVWƗQƯ'ƝQƯJ)HVWVFKULIWIRU0DKHPRXG-DDIDUL'HKDJKL: 183–202. Tehran.
Korn, A. (2005). Towards a Historical Grammar of Balochi. Studies in Historical PhoQRORJ\DQG9RFDEXODU\ (%HLWUlJH=XU,UDQLVWLN :LHVEDGHQ
Kulikov, L. (2011). “Drifting between Passive and Anticausative. True and Alleged Accent Shifts in the History of Vedic-Ya-Presents”, in: -RXUQDORI/DQJXDJH5HODWLRQVKLS. Piscataways. Available Online: https://www.degruyter.com/document/
doi/10.31826/9781463234119-013/html.
0DF.HQ]LH '1 ³1RWHV RQ WKH 7UDQVFULSWLRQ RI 3DKODYL´ LQ Bulletin of the
School of Oriental and African Studies 30: 17–29.
0DF.HQ]LH'1 $&RQFLVH3DKODYL'LFWLRQDU\. London.
0DF.HQ]LH'1 ³)UDKDQJƯ3DKODZƯJ´LQ(QF\FORSDHGLD,UDQLFD;±
1\EHUJ+6 $0DQXDORI3DKODYLPart II:*ORVVDU\:LHVEDGHQ
1\EHUJ+6 %8WDV )UDKDQJL3DKODYƯN(GLWHGZLWK7UDQVOLWHUDWLRQ7UDQVFULSWLRQDQG&RPPHQWDU\IURPWKH3RVWKXPRXV3DSHUVRI+HQULN6DPXHO1\EHUJE\%R8WDVZLWKWKH&ROODERUDWLRQRI&KULVWRSKHU7ROO:LHVEDGHQ
21
Estudios Iranios y Turanios
1\EHUJ+6 $0DQXDORI3DKODYLPart II.*ORVVDU\:LHVEDGHQ
Pirart, E. (1992). .D\kQ<DVQ <DVKW /¶RULJLQH$YHVWLTXH'HV'\QDVWLHV0\WKLTXHVG¶,UDQ. Vol. 2 (Aula Orientalis-Supplementa). Sabadell (Barcelona).
6KDNHG6 ³$'LFWLRQDU\RI$UDPDLF,GHRJUDPVLQ3DKODYL5HYLHZRIǥ)UDKDQJL
3DKODYƯN¶E\%R8WDVDQG+HQULN6DPXHO1\EHUJ´LQ-RXUQDORIWKH$PHULFDQ
2ULHQWDO6RFLHW\ 113 (1): 75–81.
6LPV:LOOLDPV 1 ³1RWHV RQ 0DQLFKDHDQ 0LGGOH 3HUVLDQ 0RUSKRORJ\´ LQ
Studia Iranica 10: 165–76.
6LPV:LOOLDPV 1 Bactrian documents from northern Afghanistan (Studies in
the Khalili collection v. 3, pt 2, 6). Oxford-New York.
7HGHVFR3 ³'LDOHNWRORJLH'HU:HVWLUDQLVFKHQ7XUIDQWH[WH´LQ0RQGH2ULHQWDOH
15: 184–258.
:HVWHUJDDUG1/ Zendavesta, or The Religious Books of the Zoroastrians. Copenhagen.
22