AI57 Cantera

This pdf is a digital offprint of your contribution in E.
Pirart (ed.), Etudes de linguistique iranienne in memoriam

Xavier Tremblay (Acta Iranica 57), ISBN 978-90-429-3264-
7
The copyright on this publication belongs to Peeters

Publishers.
As author you are licensed to make printed copies of the

pdf or to send the unaltered pdf file to up to 50 relations.
You may not publish this pdf on the World Wide Web –
including websites such as academia.edu and open-access
repositories – until three years after publication. Please
ensure that anyone receiving an offprint from you
observes these rules as well.
If you wish to publish your article immediately on open-

access sites, please contact the publisher with regard to
the payment of the article processing fee.
For queries about offprints, copyright and republication

of your article, please contact the publisher via
peeters@peeters-leuven.be
Acta iranica
acta iranica 57
études de linguistique iranienne

in memoriam
xavier tremblay
Contributions rassemblées par
éric Pirart
in aedibus PEETERS
peeters LEUVEN - PARIS - BRISTOL, CT
lovanii 2016
98106_Acta Iranica_57_vw.indd 3 24/05/16 13:39

table des matières
Éric Pirart
Préface . . . . . . . . . . . . . . . . . . . . . .VII
Philippe Swennen
Xavier Tremblay et la liturgie longue proto-indo-iranienne. Présen-
tation ....................... 1
Annexe I ....................... 19
Annexe II...................... 29
Alberto Cantera
On Avestan text criticism (2): the accusative singular of the ū- and
a- stems in the long liturgy............... 89
Juan José Ferrer Losilla
Preconsonantal nasals in the Avestan alphabet......... 157
Jost Gippert
Albano-Iranica II: Avestan +afše.............. 185
Jean Kellens
Deux apologues sur le feu rituel............... 195
Jaime Martínez-Porro
The orthography of the Avestan diphthongs aē and aō in the manu-
scripts of the long liturgy................ 205
Antonio Panaino
The World’s Conflagration and the Manichaean “Great Fire” of
1468 years...................... 221
Éric Pirart
Les cvi de l’Avesta.................... 241
Nicholas Sims-Williams
Bactria—Balkh: variations on a place-name.......... 273
98106_Acta Iranica_57_vw.indd 5 24/05/16 13:39

On Avestan text criticism (2):
the accusative singular of the ū- and ṷa- stems
in the long liturgy1
Alberto Cantera
Abstract
According to description in grammars there is in Avestan a vacillation between
‑um and ‑ūm as ending for the acc.sg. of the u-, ū- and ṷa- stems. According to
De Vaan (2003) this vacillation reflects just transmission errors in the manu-
scripts. The universal ending was in the he supposed (Sasanian) archetype
-ūm for the acc.sg of these stems, with the only exception of the ŋvha‑ stems
(< ºsṷa-), whose ending was ‑ŋvhəm. In this paper I show that the distribution of
1
This paper is part of the Avestan Digital Archive research project: Editing the Zoro-
astrian Long Liturgy (EZLL) (FF 2014-52438-P). It belongs to a series of papers on
Avestan text criticism, the first of which (“On Avestan text criticism (1): the importance
of the Iranian manuscripts for an edition of the long liturgy”) will be published in Schmitt,
Rüdiger-Sadovski, Velizar-Luschützky, Hans Christian (eds.): Iranische und indo ger
manische Sprachwissenschaft und Onomastik. Akten des Symposiums in memoriam
Manfred Mayrhofer (1926-2011), organisiert vom Institut für Iranistik und der Wiener
Sprachgesellschaft (Wien, 10.-12. Mai 2012). Wien: Verlag der ÖAW. (Sitzungsberichte
der ÖAW / Iranistische Onomastik, Nr. 14).
In the present paper I quote only readings that have been checked directly in manu-
scripts already published in facsimile or available in the Avestan Digital Archive. I shall
not use Geldner’s sigla for the designation of the manuscripts, since many of these manu-
scripts were not used by him. The habit of quoting manuscripts according to their signa-
tures in the host libraries or according to other ad hoc criteria for the manuscripts not used
by Geldner leads to much confusion. I have therefore given new sigla to all manuscripts
of the long liturgy. They are numerical sigla that give information about the text type of
the manuscript and its relative chronology:
Liturgical Combined Exegetical
Iranian Indian Pahlavi Sanskrit
Yasna 1-100 100-299 400-499 500-649 650-699
Yasna ī Rapihwin 300-399
Wisperad 2000-2199 2300-2599 2600-2899
Widēwdād 4000-4199 4200-4599 4600 4750-4760
Wištāsp Yašt 5000-5199 5200-5299 5300-5399
In each class, the manuscripts are arranged chronologically where the date is known.
A complete concordance of sigla can be found at Cantera 2014: 403 ff.
Book 1.indb 89 24/05/16 13:44

90 a. cantera
these endings is different in the Indian and in the Iranian manuscripts. In the
latter, the acc.sg. of the u- and ū- stems (-ūm viz. –īm) is still distinguished from
then acc.sg. of ṷa- (-um after single consonant [including –ŋvhum] and ‑ūm after
two consonants [including ‑ŋvhūm]). In the Indian manuscripts the lengthening of
–um into –ūm eliminated, however, this distinction. A further phonetic evolution
peculiar to the Indian transmission is the evolution of –oiium ( <*-aiṷam) to –ōim
in bisyllables (but –ōiiūm in drisyllables). The discovery of the different regular
treatment of these endings in the Indian and Iranian transmission has broad
implications. The recitation of the Avestan recitatives kept changing after the
emigration of a part of the community to India and our manuscritpts reflect the
peculiarities of both ways of reciting. Accordingly, variant readings are not just
corrupt readings of the supposed Sasanian archetype, but real recited forms when
the manuscripts were copied. We must redefine accordingly our goals when edit-
ing the Avestan texts and describing its grammar and historical phonetics.
Keywords: Avestan grammar, Avestan transmission, Avestan manuscripts
In Hoffmann-Forssman’s (1996) description of the forms of accusa-

tive singular of several stems (-u, -ū and -uua) we find a vacillation
between -um and -ūm, whose distributional rules do not appear to be
clear. The ending of the accusative singular of the u- stems is -ūm for
YAv., but for OAv. -um (for xratum) and ‑ūm (ahūm, daxiiūm). For the
ū- stems, three different forms are recorded: in YAv. -um (YAv. yauuaē-
sum) and ‑ūm (YAv. tanūm) and in OAv. the same ‑ūm (fsəratūm, tanūm)
and ‑uuəm (disyllabic, OAv. tanuuəm, cf. Ved. tanuvàm). Furthermore,
for the ṷa- stems Hoffmann-Forssman seem to consider ‑um as the usual
ending (YAv. haurum, ōiium), but ‑ūm before °ca (hōiiūmca beside
hōiium). They mention ˟ōim as well as ōiium.
In this description they just put together the forms as they were edited
by Geldner (with the exception of OAv. xratum), without any indication
of vacillation in the manuscripts. In fact, Geldner edits the ending of the
acc.sg. of the u- stems mostly as -ūm, and as ‑um for the same form of
the ua- stems (but ‑ūm before °ca, hōiiūmca Y11.4, 4, Yt15.131). For
the ū- stems we find in Geldner’s edition ‑ūm (tanūm Y10.14, 11.10,
68.23, etc.) as well as ‑um (only for yauuaē.sum: Yt19.11, 89; zauuanō.
sum Yt19.52). Hintze tries to eliminate this double representation of the
acc.sg. ending of the ū- stems by editing zauuanō.sūm and yauuaē.sūm
against all manuscripts and without indicating in any way that hers is a
correction of the transmitted text.
In his monumental work on the Avestan vowels De Vaan dedicates
the whole section 12 (2003: 319-326) to this problem. According to
Book 1.indb 90 24/05/16 13:44

on avestan text criticism (2) 91
him, Geldner edits ‑ūm for the acc.sg. of the u- and ū- stems2, but ‑um of
the ṷa- stems, but he asserts (2003: 319):
“It can be shown that the reflex of *-ṷam behaves identically to that
of *-ūm, both endings yielding ‑ūm. Thus, the ending ‑ūm is completely
parallel to the ending ‑īm. The only exception is formed by the subgroup
of forms ending in *hṷam, reflected as ‑ŋhum in our texts which was
probably rendered by *-ŋvhəm in the archetype”.
In the following I will try to show that De Vaan’s conclusions are a
simplification. Different classes of manuscripts make for a different
treatment of these endings and it is not always advisable to reduce them
to a hypothetic archetypal form. Besides, at least some classes of manu-
scripts distinguish perfectly between the endings of the ū-stems and
those of the ua-stems. Therefore, it needs to be investigated whether the
confusion of all these -ūm endings in some classes of manuscripts is the
original situation, as De Vaan supposes, or whether it is a peculiarity of
the transmission and recitation of the Avestan texts in some times or
areas. In fact, as I will show in this paper, the manuscripts do not repro-
duce the spelling of the Sasanian archetype with some deviations or
transmission errors that corrupt the Sasanian original. The way how the
Avestan texts were recited has kept changing until the present times. The
manuscripts reflect these changes at least partially in the recitation. One
of the most obvious differences is the way in which the Avestan texts
were recited in India and in Iran. Such differences have left traces in the
manuscripts. Therefore, every textual-critical research on the Avestan
text must analyse separately the Indian and the Iranian evidence, first,
and then, wherever possible, even minor regional or temporal variants.
Some of the vacillations we find in our editions are vacillations
between the Indian and the Iranian spellings. In this paper we will dis-
cover a good example in the vacillation between the Iranian ōiium and
the Indian ōim as acc.sg. of aēua- “one”. Furthermore, it seems that the
Iranian manuscripts show a different distribution of the final ‑ūm/-um
and probably a more archaic one than the Indian manuscripts. As a mat-
ter of fact, Iranian manuscripts often seem to preserve more archaic fea-
tures. A similar case is the distribution of ‑uš/ūš. Whereas in the Iranian
manuscripts ‑uš is used for the n.sg. of the u- stems and ‑ūš for the acc.
pl., as would be expected from a historical point of view, the Indian
manuscripts have generalised the ending ‑ūš for both (De Vaan 2003:
2
He neither mentions nor discusses zauuanō.sum and yauuaē.sum.
Book 1.indb 91 24/05/16 13:44

92 a. cantera
327). For further examples of the more archaic features of the Iranian
manuscripts s. (Cantera (in press)).
Until recently such an investigation was quite difficult, if not impos-
sible. Geldner does not quote in his edition all attested variants of each
edited word, but only relevant readings, so “as to allow a judgement”
(Geldner 1886: lii; Cantera 2012b: 469). Besides, when a divergent
reading is quoted, only the first manuscript mentioned by him contains
exactly this reading, while the rest of the manuscripts mentioned can
show slight variations. As a result, on the basis of Geldner’s edition it is
impossible to know the variants attested in the manuscripts of each sin-
gle word. We have only a simplified selection of variants and even these
are grouped without taking into account minor differences. Geldner’s
edition of the Avesta has filtered the Avestan transmission. Therefore,
only the direct usage of the manuscripts allows us to know the richness
and complexity of the attested variant readings.
I have conceived the Avestan Digital Archive (ADA)3 as a tool for
solving this limitation of Geldner’s edition. On this web site we publish
Avestan manuscripts indexed and with metadata that allow the compari-
son of a selected passage in all published manuscripts. In the first phase
we have concentrated our efforts on the manuscripts of the Avestan long
liturgy. Geldner used around 70 manuscripts of the different variants of
this liturgy. In ADA 98 manuscripts have already been digitised and 64
manuscripts have been published online. And this number keeps con-
stantly increasing. Today, it is already possible to make a research about
the orthographic conventions of the different manuscripts and manu-
script classes on the basis of the manuscripts published in ADA, inde-
pendently of Geldner’s edition.
This paper is one of the first attempts to review Geldner’s and subse-
quent editorial decisions on the basis of the manuscripts published in
ADA. I have checked the forms in question in the long liturgy (Yasna,
Wisperad, Widēwdād) in the manuscripts that have by now (February
2013) been totally or partially published in the ADA. Additionally,
Mihrābān’s manuscripts have been checked, although only 4600 (L4)
has been published, because of their importance and of the role they
have played in former editions of the Avestan texts4. The forms analysed
http://www.avesta-archive.com or http://ada.usal.es.
3
The manuscripts checked are the following:

4
– Iranian Sādes: Yasna 20; Wisperad 2005, 2010, 2101, 2102, 2104; Widewdad
4000, 4010, 4025, 4040, 4055, 4055, 4060, 4100, 4140
Book 1.indb 92 24/05/16 13:44

are not all attested acc.sg. of the ū- stems, but a representative selection.
For other phenomena less well attested I have checked all the forms.
When a form was frequently attested, I have analysed only a selection of
passages trying to cover the different texts and manuscript classes, but
only when the results obtained were basically uniform.
The accusative singular of the u- stems

The acc.sg. of the non-diphthongal u- stems is according to Hoff-
mann-Forssman (1996: 130) ‑ūm in Young Avestan, but ‑ūm and ‑um in
Old Avestan. The ending ‑um is only quoted in xratum. Actually, the
acc.sg. of xratu- is attested twice in the Old Avesta. In both passages
(Y28.1 and Y32.9) there is (among the manuscripts checked) not a sin-
gle manuscript with the ending ‑um, nor does Geldner quote any manu-
script with this ending. The only attested variant readings are the usual
variants for -ūm in Old and Young Avestan. From the middle of the 17th
century on, the Iranian manuscripts regularly show ‑īm instead ‑ūm:
Y28.1 xratīm: 20, 4050, 4055; Y32.9 xratīm: 20, 2101, 4025, 4050,
4055. The Indian manuscripts (including Mihrābān’s) occasionally have
a final ‑ū instead of ‑ūm: Y28.1 xratū: 230; Y32.9 xaratū: 4510, xratū:
4515. We can therefore safely delete the form xratum from the forms of
the acc.sg. of the u- stems.
Although this ending appears mostly as ‑ūm, we occasionally find ‑um
in the manuscripts. In De Vaan’s opinion (2003: 319 f.), the fluctuation
‑ūm /-um in the acc.sg. of the u- stem is the same as in the acc.sg. of the
ṷa- stems. He attributes the readings with ‑um to the Iranian Wīdēwdād
Sade branch (the manuscripts Mf2, Jp1 and Pt4). Thus the instances in
which Mf1 has ‑um are attributed to the influence of the manuscripts of
this branch5. Nevertheless, in our sample Y9.21 is the only instance of
– Indian Sādes: Yasna: 100, 120, 230, 231, 234, 235; Wīdēwdād 4200, 4210, 4230,
4240, 4250, 4400, 4420, 4510, 4515
– Combined Yasna: 400, 415, 420
– Yasna Pahlavi: 500 (J2), 510 (K5), 530
– Yasna Sanskrit: 677 (S1), 680, 681, 682
– Wīdēwdād Pahlavi: 4600 (L4), 4610 (K1), 4660, 4670, 4700, 4710, 4711, 4713,
4715.
This is a total of 49 manuscripts. In order to check that my results have not been condi-
tioned by the selection of the manuscripts, I have also used 98 complete transliterations of
Y9 recently produced by the ADA team. We have checked the relevant forms in Y9 and the
results were always largely concurring with the results obtained in the other samples.
5
De Vaan repeats Geldner’s view on the position of Mf1, according to which it is a
Sade extracted from a Pahlavi manuscript like Pt4 or Mf4. Yet there is no evidence that
Book 1.indb 93 24/05/16 13:44

94 a. cantera
‑um for an acc.sg. of the u- stems in the Iranian Sādes (s. table 1a). For
this exception cf. below.
In fact, the instances of ‑um instead of ‑ūm quoted by De Vaan are all
problematic. In Vr2.7 and 14.3 ahūmca ratūmca appears in a kind of
gloss or commentary: hō zī asti ahūmca ratūmca yō ahurō mazdå
(so Geldner). Here the acc.sg. ahūmca ratūmca make no sense at all.
We should rather read the passage as hō zī asti ahumaca ratumaca yō
ahurō mazdå “in fact, it is Ahura Mazdā who owns the ahu and
the ratu”. Thus it is already edited by Kellens (1996: 97; 2006: 43).
ahumaca ratumaca are most likely n.sg. of ahuman- ratuman-6. The
manuscripts clearly support this reading, Vr2.77:
– ahumaca ratumaca: 2010, 2101, 2102, 2104, 2230, 4000, 4010, 4025,
4040, 4050, 4055, 4060, 4100, 4140
– ahumaca ratūmca: 4200
– ahūmca ratūmca: 4210, 4230, 4240, 4250, 4400, 4410, 4510, 4515
– ahumca ratūmca: 4420
The reading ahūmca ratūmca is a lectio facilior common only in India.

Y9.21 tāiium/tāiiūm is a difficult case. The acc.sg. tāiiūm/tāiium is
attested twice: Y9.21 and Y11.3. In Y9.21, with the exception of the
Indian Sādes (which show tāiiūm), tāiium is, indeed, the best attested
reading in the rest of the manuscript classes:
1. tāiium: [Iranian Sādes] 15, 20, 82, 83, 2010, 2040, 2101, 2104,
4000, 4010, 4020, 4025, 4040, 4045, 4050, 4055, 4060, 4070, 4100,
4140, 5020; [combined Yasna] 408, 409, 410, 415, 420, 450; [Pahlavi
Yasna] 500, 602; [Sanskrit Yasna] 677
2. tāiiūm: [Indian Sādes] 110, 120, 234, 235, 4200, 4210, 4220, 4240,
4250, 4340, 4360, 4410, 4420, 4440, 4450, 4500, 4503, 4504, 4506,
4507, 4510, 4515; [combined Yasna] 400; [Pahlavi Yasna] 510, 530,
605, 613; [Sanskrit Yasna] 682
Mf1 is extracted from an old Pahlavi Yasna manuscript . It is rather an Indian copy of a
real Iranian Yasna Sade and hence not influenced by the Iranian Sāde manuscripts but
probably itself one of them (Cantera 2013).
6
Perhaps even a wrong n.sg. of ahumant- ratumant-.
7
A similar distribution appears in Vr14.3.
Book 1.indb 94 24/05/16 13:44

Nevertheless, in Y11.3 not a single variant with ‑um is found. Together

with mourum (cf. below), this form is the only acc.sg. of a u- stem in
which a variant reading with ‑um is attested in the Iranian Sādes. In the
rest of the manuscript classes, readings with ‑um are extremely rare,
occurring only in Indian manuscripts, never in the Iranian ones. But
even in the Indian manuscripts, these readings are frequent only in the
case of daxiium alongside daxiiūm. In this case, we find readings with
‑um in all passages (Y62.5, 68.5, V8.3, 18.52), although they are in the
minority: daxiiumca 682 in Y62.5; daxiiumca 120 (daxuiiumca), 234,
235, 4200, 4510 in Y68.5; daxiiumca 4200, 4210, 4230, 4240, 4515,
4601 in V8.103; °daxiium: 4670, 4700, 4712, 4713 in V18.52. Readings
with ‑ūm are, however, well attested in all passages. Strikingly, both
forms in which the variant with ‑um is more or less well represented are
forms in ºiium. This sporadic Indian spelling can be related to the Indian
reduction of (º)ōiium to (º)ōim, which is regular in disyllables and very
irregular in polysyllables (cf. below). In fact, in the Indian Sādes
gaodaiiūm (Vr2.11), although it does never appear as gaodaiium, shows
some readings comparable with ōim, hōim: gaoδāīm: 4230, gaoδāəīm:
4240, guδāīm: 4400. The Iranian manuscripts sometimes show these
words with an ending ‑əm that is exclusive of these forms: Vr2.11
gaōdāiiəm: 2104. 62.5 daxiiəmca: 20, 2101, 4050.
In fact, the acc.sg. of the u- stems is spelt almost always with ‑ūm in
all manuscript classes. Since readings with -um are relatively frequent
only in the Pahlavi Wīdēwdād manuscripts (4600 [L4] and 4610 [K1])
(cf. table 1f) but not in the Pahlavi Yasna (500 [J2] and 510 [K5])
(cf. table 1d), the oscillation between ‑um/-ūm could be a feature of
Mihrābān’s original for 4600 and 4610. The few forms with ‑um that we
find in Mihrābān’s Pahlavi Widwewdad manuscripts and their copies
are: V15.43 gātum: 4600, 4700, 4710; V13.32a pasūm 4600, 4610,
4660, 4700, 4711, 4710; V13.32b; V13.33a pasum: 4700; V13.33b
pasum: 4700; V13.34 pasum: 4610, 4711, 4715; V15.9a ratum: 4600,
4610, 4711, 4713, 4715; V15.9b ratum: 4711, 4713, 4715; V1.15
°zaṇtum: 4713, 4700; V8.10 zaṇtumca: 4601, 4713; V8.103 daxiiumca:
4601; V18.52 °daxiium: 4670, 4700, 4712, 4713.
The only real exception is mōurum, acc.sg. of mōuru- “Margiana”,
OP. margu- (N.sg. marguš, acc.sg. margum). In V1.5 there is no a read-
ing with ‑ūm. All manuscripts (including the Iranian ones) show ‑um but
some Indian Sādes attest mōurəm (4200, 4230, 4240, 4250, 4510, 4515).
In Yt10.14 it appears t5oo as mōurum. Geldner does not indicate
any variants and the manuscripts F1and E1 have in fact this reading.
Book 1.indb 95 24/05/16 13:44

96 a. cantera
Av. mōurum has obviously an ending ‑um and not ‑ūm. We can only
speculate about the reasons.
The most frequent variant of this ending is not ‑um, but ‑īm (a reading
almost totally absent from the ṷa- stems). It is the most frequent ending
in the Iranian Sādes (cf. table 1a). As is well known, in the Iranian pro-
nunciation of the Avestan, ū and ī converged in ī. This is reflected in the
manuscripts: from the middle of the 17th century, they use mostly one
single letter for ī8. However, the manuscripts of the brothers Frēdōn and
Wahrom Marzbān (2005, 2010, 4000, 4010, 4040) and also of Frēdōn
Gōbadšāh Frēdōn (4025) do still distinguish between the two sounds,
even though confusions are frequent9. Therefore, the Iranian manuscripts
copied after 1640 have always the ending ‑īm instead of ‑ūm10. In earlier
manuscripts the confusion happens occasionally. Accordingly, ‑ūm was
a real pronunciation of the ending and not just a graphical phenomenon.
Furthermore, the evolution of ‑ūm into ‑īm prevented the confusion
between ‑ūm and ‑um in the Iranian manuscripts.
Arranged by frequency, the rest of the attested variants are:
1. -ū instead ‑ūm appears almost exclusively in the Indian manu-

scripts. In Iranian manuscripts this reading can be found only for
maniiūm, the most frequent instance of the ending ‑ū for ‑ūm: Y18.8
ºmainiiū: 120, 234, 235, 4200, 4210, 4230, 4240, 4250, 4420, 4410,
4515; ºmainaiiū: 231; ºmainū: 4400; ºmainiiū 400; ºmainiiū: 415, 420;
mainiiū: 682; Y47.7 ºmainiiū: 230, 100, 120, 235, 234, 4200, 4210,
4420, 4410, 4510, 4515 ºminaiiū: 231; Vr2.8 ºmainiiū: 4240, 4410;
V10.15 mainiiū: 4000, 4010 mainiiī: 4050, 4055. It might be explained
as the result of an abbreviation that was restored wrongly to maniiū,
8
This can be the old letter for ū or, more frequently, for ī. On this confusion cf.
Martínez Porro (2012)
9
Thus we find not only –īm for –ūm, but also –ūm for –īm, e.g. Y9.8 ažūm in
4010,4025. Xōsrō Anōšagruwān Rōstām, the copyist of 4020 [Mf2] has still two different
letters, but ū is almost never used in the right positions.
10
The change of the ending –ūm into īm does not affect Mihrābān’s manuscripts.
They are older than this process. The combined manuscripts, copied in India from Iranian
originals, are also free of this confusion. I have been able to find only one instance of an
ending –īm: Y9.11 patīm [combined Yasna] 400, 450, [Pahlavi Yasna] 613 (paitīm),
[Sanskrit Yasna] 682, [Indian Sāde] 4504. Hence we can assume that their Iranian origi-
nals were copied before the 17th century. In fact, the last Iranian copy mentioned in the
colophons (Hošang Syawaxš ahryār’s) is dated 1495. The Indian manuscripts are, of
course, free from this evolution.
Book 1.indb 96 24/05/16 13:44

cf. Vr2.8 ºmainiiūš: 4400 ºmainiiəuš: 451511. There are other instances
of ‑ū for acc.sg. ‑ūm: Y9.19 ahū: 510, 530; Y11.13 tāiiū: 231; Y27.4
ahū: 682; Y28.1 xratū: 230; Y32.9 xaratū: 4510, xratū: 4515; Y62.5
ºgātū: 500; V1.15 ºjaṇtū: 4400. These readings are more sporadic than
in the case of maniiūm and limited to the Indian manuscripts, so that
maniiū can most probably be considered an abbreviation.
2. ‑ąm instead of ‑ūm is found three times in Mihrābān’s Pahlavi
Wīdēwdād (4600, 4610) and their copies: V8.40 barəšnąm 4610,
barəšnū.ąm: 4700; V15.43 gātąm: 4610, 4711, 4715; V19.29 ºpərətąm:
4600, 4610, 4710, 4711, 4715. Other than in the Pahlavi Wīdēwdād
manuscripts, we find it only in Vr15.43 gāθąm 4055 and V8.103
zaṇtąmca 4420. This fluctuation is much more frequent with the ending
‑um (cf. below).
3. The change of ‑ūm into -əm is very sporadic: Y9.11 paitəm: 415,
4210,4240,4250 patəm: 4200; V1.15 ºzaṇtəm: 4601.
4. The change of ‑ūm into -om occurs rarely: Vr2.8 ºmain̨iiom. It is
almost limited to the manuscript 4410 and the word daxiiūm; Y68.5
dāxiiaomca: 4410; V8.103 dāxiiaomca: 4410; V18.52 dāxiiaom: 4410.
5. The change of ‑ūm into ‑uum is of a purely graphical nature, analo-
gous to the fluctuation between ī, ū and ii, uu in some Indian Sādes:
Y2.2 ratuum: 682; Y9.21 tāīuum: 100, 230; Y19.6 pəratuum.cit: 231;
Y31.2 ratuum: 100; Y68.5 daxvauumca: 100.
We can conclude then that the ending of the u- stems is clearly ‑ūm for
all attested stems but mōuru- “Margiana”. In Mihrābān’s Pahlavi Yasna
manuscripts, there is almost no reading with ‑um and not a single one
exists in the Iranian manuscripts (but Y9.21 tāiium and mōurum). Other
variant readings are mostly specific of the Indian or the Iranian transmis-
sion and reflect the idiosyncrasies of the recitation in these areas.
The accusative singular of the ū- stems

The are only a few ū- stems attested in acc.sg. in the long liturgy:
YAv. and OAv. tanūm (Y33.10) (tanū “body”) alongside the trisyllabic
form OAv. tanuuəm (Y46.10), and YAv. xšnūm (xšnū- “friendly recep-
tion”). OAv. and YAv. fsəratūm could belong to a stem fsəratu- as well
11
A similar explanation was advanced by Hoffmann (1975-1992: 600, n. 15) for
Yt5.102 gātu. On the abbreviations in general cf. now Tremblay (2012: 118 ff.).
Book 1.indb 97 24/05/16 13:44

98 a. cantera
as to fsəratū-. The ending appears as ‑ūm in all manuscript classes,

exactly like the ending of the u- stems, and shows the same variant read-
ings (e.g. -īm in the Iranian Sādes) (cf. table 2). Only xšnūm is problem-
atic. This word is almost always attested with the ending ‑əm in the
Iranian as well as in the Indian manuscripts (Y48.12 [Iranian Sādes]
2010, 4000, 4010, 4050, 4055, [Indian Sādes] 120, 234, 235, 4200,
4210, 4240, [combined Yasna] 415; [Pahlavi Yasna] 500; Y53.2 [Ira-
nian Sādes] 2101, [Indian Sādes] 230, 100, 120, 234, 235, 4210, 4240,
4420, [combined Yasna] 400, [Sanskrit Yasna] 680) or -əm (Y48.12
[Indian Sādes] 230, 100, 4410, 4420, 4510, 4515, [combined Yasna]
400, 420, [Sanskrit Yasna] 682; Y53.2 [Indian Sādes] 4400, 4410, 4510,
4515, [combined Yasna] 415, 420, [Sanskrit Yasna] 682). The ending
‑ūm appears rarely, but often in the oldest manuscripts of each class:
Y48.12 [Iranian Sādes] 2005, 4000, 4010, [Indian Sādes] 231, 4400,
[Pahlavi Yasna] 530; [Sanskrit Yasna] 680; Y53.2 [Iranian Sādes] 2005,
2010, [Pahlavi Yasna] 500, 530).
The accusative singular of the ṷa- stems

De Vaan (2003: 323) writes:
“Most of the Avestan forms containing *-ṷam were edited with ‑um
by Geldner, but it appears from the evidence that, after all consonants
except *-h-, final ‑ṷam yielded ‑ūm in the archetype. Wherever -um is
philologically better attested, the surrounding forms have influenced
their spelling. The only real exception is the ending *-ahṷam- which was
spelled ºaŋvhəm in the archetype.”
De Vaan assumes that some forms like vidōiiūm present the same
distribution between ‑ūm/‑um as the acc.sg. of the ū- stems. Further-
more, for the numerous forms in which ‑um is better attested than ‑ūm
he presents various ad hoc explanations (2003: 324). For haurum the
usual ending is ‑um, but the Indian Sādes would preserve the original
‑ūm. This is an odd argument, given that De Vaan does not generally
consider the Indian Sādes very authoritative. Av. pourum would be
influenced by the frequent pouru. Other cases are influenced by sur-
rounding acc.sg. in -əm, hence the variant readings with ‑əm/-um. The
fraction nouns θrišum, caθrušum are analogical to paŋtaŋhum, which in
turn replaces the form of the archetype *paŋtaŋvhəm (!). As a matter of
fact, an analysis of the really attested forms and their distribution in the
different manuscript classes shows a totally different picture and make
all these ad hoc assumptions unnecessary.
Book 1.indb 98 24/05/16 13:44

The Iranian manuscripts perfectly distinguish between the acc.sg. of

the ū- stems with the ending ‑ūm/-īm and the acc.sg. of the ṷa- stems
with the ending ‑um (cf. table 3a). This is manifest in several expres-
sions in which both forms appear together and the endings are kept
distinct in all Iranian manuscripts, e.g.: barəšnūm paiti vaɤδanəm pou-
rum “in the high front part of the head” (V8.40, 41, 9.15), hōiium žnūm
“the left knee” (8.62, 63, V9.22, 23) or hōiium ascūm (V8.64, 65, 9.23).
The acc.sg. of the u- stems appears as ‑ūm/-īm, the latter always as ‑um.
In only one instance (V9.15 paōirīm 405012) do we find a reading with
‑īm, the regular ending of the ū- stems in manuscripts after 1640, and
never ‑ūm. There are only four exceptions that show a completely
different pattern of attestation: xštūm (Y9.21, V4.19, 4.23, 5.29), srūm
(V9.14 [3x], 16.6) and dadrūm (V9.11) and drūm (V22.5). Actually
they are not exceptions, but regular forms in which the ending ‑ūm has a
different origin than the usual ‑um (cf. below).
The variant readings for the acc.sg. of the ṷa- stems in the Iranian
manuscripts are different from the variant readings for the u- stems as
well. Whereas for the latter there is almost exclusively the variant read-
ing ‑īm, for the ṷa- stems we find the following variant readings:
– the most frequent variant reading is ‑ąm: Y19.7 θrišąm (2010, 4025);
caθrušąm (2010, 4040); V6.32 θrišąm (4050); caθrušąm (4010,
4025, 4040, 4601), ciθrušąm (4050); V16.2 θrišąm (4000, 4010,
4050, 4055); caθrušąm (4000) ; V18.63a θrušąm (4000, 4010, 4055),
θrišąm (4050); V18.64b θrišąm (4000); Y11.5 hōiiąmca (4050);
V9.18a,b haōiiąm (4050). It seems that ‑um and ‑ąm were recited
similarly and could therefore be confused. Thus Mihrābān
Anōšagruwān Wahromšāh copies Y11.5 twice in 4050, writing first
hōoiiąmca and then hōiiumca. Observe that -um appears as a variant
reading of ‑ąm in some instances in the Iranian manuscripts: Y9.18
kaōiium for kaōiiąm in 2010, 4010, 4025, 4031, 4040, 4045, 4050,
4055, 4060, 4100, 4140; Y9.18 tbišuuatum for tbišuuatąm in 4020.
– The confusion between ‑um and -əm (which could, of course, be influ-
enced by the frequent acc.sg. in -əm) is less frequent. It only appears
in the forms haurum and saurum which show a peculiar treatment in
the other classes of manuscripts as well (cf. below): y19.14 haurəm
(2010, 2101, 4000, 4010, 4040, 4055, 4060, 4100), ahurəm (20, 4025,
4050); V10.9 saōrəm (4000, 4010, 4050), sōrəm (4055).
12
Obviously influenced by the corresponding feminine form paōirīm.
Book 1.indb 99 24/05/16 13:44

100 a. cantera
Thus there is no doubt that the Iranian manuscripts still distinguish

between the acc.sg. of the ū- stems (-ūm) and the acc.sg. of the ṷa- stems
(-um). Four acc.sg. of ṷa- stems consistently show, however, the ending
‑ūm in the Iranian manuscripts and also in the rest of the manuscript
classes13 (cf. table 4): xštūm (Y9.21, V4.19, 4.23, 5.29), srūm (V9.14
[3x], 16.6), dadrūm (V9.11) and drūm (V22.5). In them the ending
always follows two consonants. The Iranian manuscripts seem then to
still distinguish between the ending ‑ṷam# yielding ‑um and ‑uṷam#
(Siever’s variant of ‑ṷam) yielding ‑ūm
This evolution of ‑uṷam is confirmed by the n.sg. of the 2.p.sg. per-
sonal pronoun tūm. It derives from *tuṷam (cf. bisyllabic Av. tuuəm,
Ved. tuvàm) and accordingly always appears in the Iranian manuscripts
as tūm/tīm14. In its 10 attestations we have not found a single reading
with ‑um in the Iranian Sāde manuscripts. If the Iranian manuscripts of
the 17th century still distinguish between ºṷam# and ºuṷam#, then the
“Persian” pronunciation of every ‑ṷa- as ‑uṷa- (for this evolution
cf. Hoffmann 1979) should be posterior to the evolution of ºṷam# into
ºum and ºuṷam# into ºūm. This evolution took place after the lengthen-
ing of final ºum to ºūm. Unfortunately we cannot date this lengthening,
but in any case the Achaemenid origin of this pronunciation (supposed
by Hoffmann) is highly questionable. If the use of uu and ii for inter
mediate i and ṷ has its origin in such a pronunciation, we will have to
date the different evolution of ºṷam# and ºuṷam# before the invention of
the Avestan script.
In the Indian Sādes this ending appears in the vast majority of the
passages and manuscripts as ‑ūm (cf. table 3b), hence De Vaan’s assump-
tion that it is the form of the archetype. Nevertheless, although ‑ūm is
the most frequent ending, in some passages ‑um is still the best or even
the only attested ending:
– pourum (acc.sg. m. of pouruua-) appears thus in all passages V4.47,

8.40, 41, 58 (twice) and 9.15. The only attested variant readings are
paōirīm/paoirīm in the manuscripts 4210, 4240 and 4410 in V8.40, 41
and 9.15 and 4230 in V8.58b and 9.15. Furthermore, 4230, written by
13
The only exception is srum in Mihrābān’s Pahlavi Wīdēwdād manuscripts and, con-
sequently, later copies of these manuscripts.
14
Thus could be explained the two spellings of the secondary middle ending of the
2.p.sg. *dhṷam: OAv. °dūm would derive from *duṷam (the form of the ending after two
consonants or long vowels) and YAv. °δβəm from *dṷam (the ending after a short vowel).
Book 1.indb 100 24/05/16 13:44

Dārāb Hīrā Candā as 4210, shows a tendency towards paourəm

(V8.41, 58b). Only in V8.40 the manuscript 4400 shows paourūm.
– saurum (V10.9; acc.sg. of sauruua-, Ved. sarvá-) appears only with the
ending ‑um (4210, 4400, 44210, 4420) or -əm (4200, 4230, 4510, 4515).
– in the Yasna haurum appears always with ending ‑əm, often trans-
formed in the lectio facilior ahurəm: Y9.14 haurəm (4420), hurəm
(4400), ahurəm (100, 120, 230, 231, 234, 235, 4200, 4210, 4240,
4250, 4410, 4510, 4515); Y20.3 ahurəm (100, 120, 230, 231, 234,
235, 4200, 4210, 4240, 4250, 4410, 4510, 4515).
Although in the rest of the attested forms the almost universal ending
is ‑ūm, we occasionally come across ‑um. This ending appears only in the
oldest Indian Sades, while in the same passages more modern manu-
scripts show ‑ūm: Y19.7 ciθrušum (230, 100); V6.32 θrišum (4200,
4210); V16.2 caθrušum (4200, 4210), ciθrušum (4240). Furthermore, we
sometimes find endings that usually alternate with ‑um and not with ‑ūm.
Thus the related manuscripts 4510 and 4515 have three times ciθrušəm
(Y19.7, V6.32, 16.2). 4400 has one occurrence of ciθrušąm (V16.2).
In Mihrābān’s manuscripts the ending ‑um is still majoritarian (cf.
tables 3d, f) in all forms, but in the acc.sg. of ºōiiūm we find exclusively
‑ūm (cf. tables 4d, f). In the rest of the forms, apart from ‑um we find the
variant reading ‑ąm which is specific for ‑um in the Iranian manuscripts.
It appears only in Mihrābān’s Pahlavi Wīdēwdād manuscripts: V16.2
θrišąm (4600), V18.63a θrišąm (4600, 4610), V18.63b θrišąm (4600);
V6.32 caθrušąm (4610), ciθrušąm (4700). Furthermore, in V6.32 both
4700 and 4713, copies of 4600, presume θrišąm in the lost 4600.
Although Mihrābān copied both manuscripts from the same source, the
readings ‑um and ‑ąm alternate in the same passage in both manuscripts.
Readings with ‑ūm appear only rarely, but are especially frequent in
4610 (K1). We find θrišūm in 510 (Y19.7) and in 4610 (V18.63b,
18.64b). In all cases the older manuscripts 500 (J2) and 4600 (K1) show
the older endings ‑um or ‑ąm. The ending ‑ūm seems to be an Indian
influence. In the combined Yasna manuscripts (400, 415, 420) the domi
nant ending is clearly ‑um (cf. tables 3e, 4e). It is also the only form I
find in 677 (S1): Y19.7 θrišum (677, 682), θrišum (680); caθrušum
(677), ciθrušum (680), ciθrušum (682), Y2.13 [vīdōii]um (677), Y11.4
hoiiumca (682).
The ending ‑um appears clearly as the older form of the ending for the
acc.sg. of the ṷa- stems. Its almost universal spelling as -ūm in the Indian
manuscripts must be the result of a secondary lengthening. It appears
Book 1.indb 101 24/05/16 13:44

102 a. cantera
occasionally as early as in Mihrābān’s manuscripts, especially in the

later ones, so that this lengthening must have taken place before the 14th
century. Interestingly, 677 (S1) seems not to be affected by it. It is only
through this secondary lengthening of ‑um into ‑ūm that the endings of
the acc.sg. of the ū- and ṷa- stems became indistinguishable in the Indian
manuscripts.
The stems ending in ºaiṷa- require special attention (cf. table 5). In
the Indian manuscripts, they present a clear distribution that has not yet
been recognised (cf. table 5b). In the Indian Sāde and in Mihrābān’s
manuscripts trisyllabic (or longer) stems show the ending ‑ōiūm with
‑ūm (whereas in the Iranian sources we find ‑um): vīdōiiūm (acc.sg. of
vīdaēuua- “exorcising the daēuua”) and harōiiūm (acc.sg. of haraēuua-
“Haraiva, a country name”). Disyllables show, however, a different
spelling. In the Indian Sāde manuscripts (from the oldest ones on), the
ending ‑ōiium15 yields ōim. As a result these disyllabic stems become
monosyllabic. Thus, in the Indian Sādes the acc.sg. of aēuua- “one” is
ōim and of haoiia- (Ved. savya-) it is hōim16. Y11.4, 5 hōiiūmca in all
Indian Sādes (with minor variants not affecting the ending) confirms that
the reduction ‑ōiium to ‑ōim affects only disyllables. The only exception
is the Indian Yasna Sāde 231, a manuscript very strongly influenced by
the modern recitation. There we find Y11.4, 5 hōīmaca. The Pahlavi
manuscripts too, including Mihrābān’s 4600 and 4610, always have ōim
and hōim (cf. tables 5d,f), but hōiiūmca Y11.4 (500, hūiiūmca 510, 530)
and 11.5 (500, 510, 530)17. In the Iranian manuscripts these acc.sg.
appear always as ōiium and hōiium, both with minimal variant readings,
but never as ōim or hōim (cf. table 4a). Unfortunately, we lack data for
the combined Yasna and for the oldest Sanskrit manuscripts.
The distribution of the forms of the acc.sg. of the stems in ‑aiṷa- can
be represented schematically as follows:
15
I assume that the reduction of -ōiium into -ōim is prior to the Indian lengthening of
-um not only because this seems more likely, but as well because the reduction affects
Mihrābān’s manuscripts, while the lengthening appears there only sporadically.
16
This is not a spelling for *haoīm, perhaps the regular acc.sg. of haoiia-, but a reduc-
tion of hōiium (attested in the Iranian Sādes) and with the enclitic ºca, hōiiūmca.
17
613 has hoiiumca in both passages.
Book 1.indb 102 24/05/16 13:44

Iranian Indian Mihrābān combined other Yasna

Sādes Sādes Yasna Pahlavi Sanskrit
manu-
scripts
aēuua- ōiium ōim ōim ------ ōim ------
(disyllables)
haoiia- + ºca hōiiumca hōiiūmca hōiiūmca hōiiumca/ hōiiūmca / hōiiumca /
hōiiūmca hōiiumca hōiiūmca
vidaēuua- vīdōiium vīdōiiūm vīdōiiūm vīdōiium/ vīdōiiūm vīdōiium/
(trisyllables or +) vīdōiiūm vīdōiiūm
Some Indian Sāde manuscripts sometimes use the ending ‑ōim (or
similars) even for trisyllables, but the expected forms are always better
attested. These variant readings belong almost always to the same manu-
scripts18, viz. such as are strongly influenced by the recitation and show
little, if any, dependence from written copies (Cantera 2012a: 301 f.):
Y2.13 vīduiim (231), vīδuiim (4400), vīdaoīm (4200); Y6.12 vīdūīm
(415, 4400); Y25.6 vīduiim (230, 100), vīδuiim (231); Y71.5 vīdōīm
(230, 120), βiduiim (231), vīdūīm (4400).
Av. daēuua- shows an unexpected acc.sg. (cf. table 6). The regular
form would be *dōiium for the Iranian manuscripts and *dōim for the
Indian ones. Instead we find almost everywhere the acc.sg. daēum (with
‑um and not ‑ūm). The diphthong aē did not yield the expected ōi because
of the influence by aē of the rest of the paradigm (daēuuō, etc.). Only
two variant readings appear: daēuum is a frequent variant reading in the
Iranian Sādes (V10.13 4000, 4050, 4055; V13.7 4000, 4010, 4050,
4055) and daem in the Indian manuscripts (already in Dārāb Hīrā Candā:
V10.9 4230, 4410; V13.7 4200, 4230) and once already in Mihrābān’s
4610 (K1) daēm19. This form is the equivalent of *dōim, for it appears
18
An exception is Y71.5 vīdōīm in Dārāb Hīrā Cāndā’s manuscript 120 (Lb2).
19
But 4600 (L4) daēum. A problem is posed by vaēm, the strange and only acc.sg. of
vaiiu- (Ved. vāyú‑). The expected form would be *voiium in Iranian manuscripts and
*vōim (monosyllabic) in the Indian manuscripts, from Mihrābān on. Actually, the only
attested form is vaēm. Unfortunately, we are not sure of the form of this acc.sg. in the
Iranian liturgical manuscripts. The only attestation in the Yasna appears in Y25.5. This
passage is part of the standard šnūman of Mīnō-Nāwār, used in the daily ceremonies. This
šnūman is usually not included in the Iranian Wisperad or Wīdēwdād Sādes. Hence,
among the manuscripts already published in ADA, this form is unfortunately only attested
in the late Yasna Sāde manuscript 20, where it appears as vaē. Geldner does not quote any
variant readings in the rest of attestations in the Khorde Avesta: S2.21, Yt15.1, Yt15.5,
Yt15.57, Y25.5. Therefore, this form is not directly comparable with the sporadic variant
reading daēm. Perhaps the original form was *vaēum parallel to daēum and not the non-
Book 1.indb 103 24/05/16 13:44

104 a. cantera
only in manuscripts with the evolution -ōiium > -ōim, but with restitu-
tion of the paradigmatic daēº instead of ºdōiº. The reading daēuum is the
result of another restitution of the stem daēuua-.
According to Hoffmann and Narten (1989: 52 n. 57) the acc.sg. of
the ṷa- stems after h, like hauuaŋvha- “good life” (< *haṷ-ṷahṷa-), was
hauuaŋvhəm in the archetype. This would be the only form that could
explain the really attested forms hauuaŋhəm and hauuaŋhum. The form
supposed for the archetype is not justified by palaeographic or textual-
critical evidence. It is just the lowest common denominator to the
attested forms and thus a reconstructed form with respect to which we
cannot know whether it was ever written down as such. The only argu-
ment for postulating it is that both forms must derive from an “arche-
typal” common form and that we suppose it to be the direct ancestor of
the different attested forms. In fact, there are no more arguments for
postulating a Sasanian *hauuaŋvhəm than a *caθrušuuəm.
De Vaan, however, thinks the confirmation of this postulate lies in the
distribution of the variant readings of paŋtaŋvha- “one fifth”. Actually,
the attested forms show almost the same distribution of the variant read-
ings as for the rest of the uua- stems. The Iranian Sādes never show a
final ºŋvhəm20, but always the two more frequent endings for the ṷa-
stems, i.e. ‑um and ‑ąm (cf. table 7a):
■ -um:
– for paŋtaŋhum: :Y19.7 paŋtaŋhum: 4000, 4010, 4055 paṇŋtaŋhum: 4060,
4100, paṇgtaŋhum: 2101; V6.32 paŋtaŋhum: 4000; V16.2 paŋtaŋhum:
4010, 4055
– for hauuaŋhum: Y71.11a hauuaŋhum: 2101, 4000, 4010, 4055;
hauuaŋvhum: 2010; Y71.b hauuaŋvhum: 2101; hauuaŋhum: 4000, 4010,
4050, 4055
– -ąm: Y19.7 paŋtaŋhąm: 4025, 4040, 4050, paṇgtaŋhąm: 20, pataŋhąm:
2010; V6.32 paṇgtaŋhąm: 4010, 4025, 4040, paŋtaŋhąm: 4050, 4055;
V16.2 paŋtaŋhąm: 4000, 4050.
paradigmatic *vōim. The attested vaēm could be the result of a dissimilation or even a
faulty restitution of one abbreviated form vaē., cf. Y25.5 vaē (20). In any case, De Vaan
(2003: 326) is right in rejecting the derivation of vaēm from an IIr. stem vāiá- (hesitat-
ingly suggested by Hoffmann and Forssman 1996: 58).
20
The only exception is 2010 Y71.11a, but it was corrected into hauuaŋhum.
Book 1.indb 104 24/05/16 13:44

In the Indian Sādes we find three endings that also appear for the
rest of ṷa- stems, but this time the distribution is not exactly the same
(cf. table 6b). Here the ending -əm is visibly more frequent than in the
rest of ṷa- stems. Nevertheless, in all passages the ending ‑um or ‑ūm is
better represented (but V16.2):
paŋtaŋvhūm -um -ūm (or ‑uum) -əm/-əm -ąm

Y19.7 paṇgtaŋhum: paṇgataṇghūm: 4200, paṇɤa.tṇɤəm: 231
100 4210, 4240 paṇgtaŋhəm: 120,
pgŋtaŋhum: pəṇgataṇghūm: 4250 234
677 paṇgataṇgūm: 4420 paṇg.taŋhəm: 235
paṇgataŋghūm: 4410 paṇgataŋhəm: 4510,
paŋhe.taŋhūm: 4400 4515
paŋhataŋhəm: 420
paŋtaŋhəm: 500, 530
V6.32 paṇgtaŋhūm: 4200, paṇgtaŋhəm: 4230 paŋtaŋhąm:
4210, 4420 paṇatatŋhəm: 4410 4711, 4715
paŋataŋhūm: 4340 paṇg.taŋhəm: 4510, paŋhtaŋhąm:
paŋhδaŋhūm: 4400 4515 4601, 4712
paŋhtaŋhəm: 4670, paŋhaitaŋhąm:
4713 4700
paŋhaotaŋhəm: 4710
V16.12 paŋhu.taŋhūm: 4400 paŋtaŋhəm: 4700 paŋtaŋhąm:
paŋhataŋhəm: 4670 4600, 4660,
paŋhutaŋhəm: 4200, 4711, 4713
4210, 4240, 4420 paŋhąm: 4710
paŋhutaŋhm: 4230
paṇŋhataŋ́həm: 4410
paṇg.taŋhəm: 4510
paṇg. taṇgəm: 4515
hauuaŋvhūm -um -ūm (or ‑uum) -əm/-əm -ąm
Y71.11a hauuaŋhum: hauuaŋhō.hūm: 231 hauuaŋhəm: 400, 415

4420 hauuaŋhūm: 4510 hauuaŋhəm: 500
huuaŋum: huuaŋhūm: 4515 hauuaŋhəm: 500,
4240 hauuaŋhauum: 100 530, 682
auuaŋhum: huuaŋuhuum: 235, 234
4410
Y71.11b hauuaŋum: hauuaŋhūm: 235, 234, hauuaŋhəm: 400,
4240 4420, 4510 415, 420
auuaŋhum: huuaŋhūm: 4515
4410 hauuaŋhauum: 100
Book 1.indb 105 24/05/16 13:44

106 a. cantera
This frequency of -um/-ūm never occurs in other final words with

final -əm. The question is whether the frequency of -əm in Mihrābān’s,
the combined Yasna manuscripts and especially in the Indian Sādes
reflects a peculiar development of the recitation affecting only a part of
the transmission (like the lengthening of ‑um# into ‑ūm#), or whether is
an old feature (the non-coloration of -əm# after ŋvh in the same way as
after uua) lost in other areas. The second explanation fits better with
Hoffmann-Narten’s assumption of an archetypal ºŋvhəm that appears
then in a part of the manuscripts as ºŋ(v)hum/ ºŋ(v)hūm and in another part
as ºŋhəm. Although we cannot rule out this explanation, it poses some
difficulties. The coloration was universal in the tradition underlying the
Iranian Sādes. The rest of the manuscripts will attest a fluctuation
between coloration and non-colaration. To attribute the coloration of ‑əm
into ‑um to the labiality of ŋvh at the time of the writing down of the
Iranian Sādes is, however, problematic, since even Mihrābān’s manu-
scripts and the oldest Iranian and Indian Sādes do not reflect the labiality
of this group either for hauuaŋvhum or for paŋtaŋvhum. As I have
recently shown (Cantera (in press)), although the use of ŋh for ŋvh is
quite frequent in all manuscripts, ŋvh is still used in the expected posi-
tions in the oldest Iranian manuscripts like 4000 and 4010. Furthermore,
the Indian Sādes, while not attesting ŋvh, quite often use ŋuh for it. In the
attestations of hauuaŋvhum and paŋtaŋvhum, ŋvh appears very rarely in
Iranian manuscripts, namely only in the manuscript 2010 at Y71.11a
(hauuaŋvhum) and in 2101 at Y71.11b (hauuaŋvhum). The alternative
ŋuh is totally absent from the Indian manuscripts. Furthermore, ‑um or
the alternative ‑ąm appear already in Mihrābān’s and in the oldest Indian
Sādes. Therefore, it is more likely that the regular ending of the acc.sg.
of the stems in ºŋvha- was ‑ŋhum. In India ºŋhum was recited probably in
a way very similar to ºŋhəm21 and this led to confusion, thanks also to
the frequency of -əm# in the acc.sg.
The development of the acc.sg. of these stems

The manuscripts show a variety of readings for the ending of the acc.
sg. of the ū- and ṷa- stems. However, an analysis of the readings attested
in each different class of manuscripts, combined with a consideration of
the geographical and chronological distribution of the manuscripts,
reveals more or less clear distributions:
21
Compare the frequent confusion between -um and -ąm in the Iranian manuscripts.
Book 1.indb 106 24/05/16 13:44

Iranian Indian Mihrābān combined other Yasna

Sādes Sādes Yasna Pahlavi Sanskrit
manu-
scripts
u- stems -ūm -ūm -ūm -ūm -ūm -ūm
-īm (-ū) (-um)
(-um)
ū- stems -ūm -ūm -ūm -ūm -ūm -ūm
-īm
ºṷa- stems -um -ūm -um -um -ūm -um
-ūm (after 2 -ąm -ūm -um (-ūm)
consonants) -ūm -ąm
trisyllabic -ōiium -ōiiūm -ōiiūm -ōiiūm -ōiiūm -ōiiūm
ºaiṷa- stems -ōiium -ōiium (-ōiium) -ōiium
disyllabic (º)ōiium (º)ōim (º)ōim ------ (º)ōim ------
ºaiṷa- stems
stems ºŋ(v)hum ºŋhūm ºŋhąm ºŋhəm ºŋhąm -ŋhum
in ºŋvha ºŋhəm ºŋhəm ºŋhəm
ºŋhum
This distribution can be explained as the result of a series of evolu-

tions of the expected ending for an Old Iranian language (-um, -ūm and
‑ṷam or -ṷəm). Some of these affected the totality of manuscripts, but
others extended only to a part of the transmission. The most widespread
phenomenon is also the oldest one: the lengthening of ‑um (acc.sg. of
the u- stems) to ‑ūm, so that the distinction between the acc.sg. of the
u- and ū- stems faded away. As the evidence of the Iranian Sādes and
isolated forms in other manuscript classes show, this lengthening did not
affect the outcome of ṷ əm. Therefore, it should have taken place before
the evolution of ‑ṷəm > -um. At a later stage, ‑ṷəm yielded -um, and
‑uṷəm (a Sievers variant of ‑ṷəm) became ‑ūm. This is exactly the situa-
tion reflected in the Iranian Sādes: the acc.sg. of the ū- stems is ‑ūm and
the acc.sg. of the ṷa- stems ‑um or ‑ūm, when the suffix follows two
consonants. This evolution happened before the pronunciation uṷV for
older ṷV was generalized, if such evolution did ever take place.
Another group of manuscripts, represented mainly by the Indian
Sādes, reflects a tradition of recitation in which further changes took
place. First, the new ‑um (< ‑ṷəm) was lengthened as well and got con-
fused with the ending of the ū- stems, so that the distinction between
both stems disappeared. The spelling ‑um continued to be used sporadi-
cally, but it was pronounced exactly like ‑ūm. Furthermore, in the same
Book 1.indb 107 24/05/16 13:44

108 a. cantera
group of manuscripts disyllabic words ending in ‑ōiium were reduced

to ‑ōim, thereby becoming monosyllabic, probably before the second
lengthening of -um.
As we see, the endings of the ū- and ṷa- stems in the Avestan manu-
scripts do not reflect how these endings were written in an imaginary
Sasanian archetype. They rather reveal how these endings were recited
in the different traditions. The Iranian manuscripts appear clearly as con-
stituting a tradition by itself, a tradition which, as so often, shows more
archaic features here than the alternative tradition represented by the
Indian Sāde manuscripts (Cantera (in press)). Mihrābān manuscripts and
the so-called “Iranian” Pahlavi Yasna held a kind of middle position
between the “Iranian” and the “Indian” tradition. The “Iranian” Pahlavi
Yasna manuscripts are in reality Indian manuscripts going back on
an Iranian original copied two centuries before the extant manuscripts.
We neither know how many copies there are in between, nor when these
copies arrived in India. In any case, Indian influence is by no means to
be ruled out.
The case of the Mihrābān manuscripts is different, for here we have
the originals (500, 510, 4600, 4610). In them we already find the reduc-
tion of ‑ōiium to ‑ōim in disyllables which does not appear in the Iranian
Sādes, but the lengthening of ‑um (< *-hṷam) is only rare. In fact, these
are not the only features that Mihrābān’s manuscripts share with the
later Indian manuscripts. Thus Mihrābān uses the “Indian” initial y22
and not the Iranian one we find in the Iranian Sādes and, alternating with
the Indian, in the oldest Indian Sādes. Furthermore, he only seldom uses
n, just like the Indian Sādes where it is totally absent, whereas its usage
is regular in the Iranian Sādes. Mihrābān therefore either represents an
alternative Iranian tradition (different from the tradition represented in
the Sādes) or, since his manuscripts were copied in India and perhaps
from originals already available in India, they might be influenced by
the Indian tradition. The position of Mihrābān’s manuscripts in the
transmission of the Avesta needs to be further researched (the same is
actually true for almost all important Avestan manuscripts and copyists).
This research on the acc.sg. of the ū- and ṷa- stems shows that the
forms attested in the manuscripts are the result of the evolution of the
pronunciation of the Avestan language after the Sasanian times. Even
22
Strikingly Tremblay (2012: 129 f.) in his discussion of the matter does not take into
consideration that Mihrābān uses this letter.
Book 1.indb 108 24/05/16 13:44

after the segregation of the Indian and Iranian communities the way how
the Avestan language was recited in the ritual continued changing and
these changes are reflected in the manuscripts. Thus, the forms attested
in the manuscripts are not degenerations of a supposed original Sasanian
archetype or a subsequent hyparchetype through the copying process23.
But even if we accept the existence of a Sasanian archetype for the
extant manuscripts, the attempt to edit the recitative of the long liturgy
in the form in which it was written down in the “liturgical” Sasanian
archetype is too risky. Returning to the acc.sg. forms we have discussed,
it is impossible to know which was the stage reached at the time of the
supposed Sasanian archetype. Had the distinction between the acc.sg. of
the u- and ū- stems already faded away? Had -ṷəm# already evolved
into ‑um and -uṷəm# into ‑ūm? Being strict, we can only edit either the
Iranian redaction (with its distinction of the acc.sg. of ū- and ṷa- stems)
or the Indian one (without such distinction and with reduction of ºōiiūm
to ºōim in disyllables). The phonetic changes of the Avestan language
did not come to an end in Sasanian times but continue until today. The
extant manuscripts are not defective copies of a Sasanian archetype but
reflect the way the Avestan text was recited in their respective traditions.
Editions of the Avestan texts must take into account this fact.
23
Indeed, the existence of the archetype and the hyparchetypes has recently been
doubted (Cantera 2012a; Tremblay 2012). In any case, even if there was a Sasanian
archetype, our manuscripts do not derive directly from it (Kellens 1998; Kellens 2012).
Book 1.indb 109 24/05/16 13:44

110 a. cantera
Tables
Table 1 : u- stems
Table 1a Iranian Sādes
ahūm -um -ūm other endings
Y9.19 ahūm: 2101 ahīm: 15, 20, 82, 83, 2010,
2040, 2104, 2801, 4000, 4010,
4020, 4025, 4031, 4040, 4045,
4050, 4055, 4060, 4070, 4100,
4140, 5020
Y11.10 ahūm: 2010, 4000 ahīm: 20, 2101, 2104, 4140
Y27.5 ahūm: 2005, 2010, ahīm: 20, 2101, 4050, 4055,
4000. 4010, 4025, 4100
4040
Y30.6 ahūm: 2005, 2010, ahīm: 20, 4040, 4050, 4055
2101, 4000. 4010,
4025
V2.22a ahūm: 4000, 4010, ahīm: 4050, 4055
4025, 4040
V2.22b ahūm: 4000, 4010, ahīm: 4050, 4055
4025, 4040
V18.16 ahūm: 4000 ahīm: 4050, 4055
barəšnūm -um -ūm other endings
V8.40 barəšnīm: 4000, 4010, 4025,
4050, 4055
V8.41 barəšnūm: 4010, barəšnīm: 4000, 4050, 4055
4025
V9.15 barəšnīm: 4000, 4010, 4055
baršnīm: 4050
dax́iiūm -um -ūm other endings
Y62.5 daxiiūmca: 2010, daxiiəmca: 20, 2101, 4050
4000, 4010, 4055
Y68.5 daxiiūmca: 2005, daxiiīmca: 20, 2101, 4050,
2010, 4000, 4010 4055
V8.103 daxiiīmca: 4000, 4010, 4025,
4050, 4055
V18.52 ºdaxiiūm: 4000, ºdaxiiīm: 4050, 4055
4010
Book 1.indb 110 24/05/16 13:44

gaōdaiiūm -um -ūm other endings

Vr2.11 gaōdāiiūm: 2010, gaōdāiiīm: 2101, 2102, 4010,
4040 4025, 4050, 4060, 4100, 4140
gaōduiiūm: 4000 gōdāiiīm: 4055
gaōdāiiəm: 2104
gātūm -um -ūm other endings
Y62.524 ºgātūm: 2005, 2010, ºgātīm: 20, 2101, 4050, 4055
4010 ºgāθīm: 4000
V9.33 ºgātūm: 4000 ºgātīm: 4010, 4050, 4055
V9.34 ºgātūm: 4000 ºgātīm: 4010, 4050, 4055
V15.43 gāθīm: 4000, 4010, 4050
gāθąm; 405525
mańiiūm -um -ūm other endings
Y18.8 ºmainiiūm: 2005, ºmainiiīm: 20, 2101, 4025,
2010, 4000, 4010, 4050, 4055, 4060, 4100
4040
Y47.7 ºmainiiūm: 2005, ºmainiiīm: 20, 2101, 4055,
2010, 4000, 4010 4055
V10.5 mainiiū: 4000, 4010
mainiiī: 4050, 4055
Vr2.8 ºmainiiūm: 2010, ºmainiiīm: 2101, 2102, 2104,
4000 4010, 4025, 4040, 4050, 4055,
4140
ºmain̨iiom: 4100
Vr19.0 ºmainiiūm: 2005, ºmainiiīm: 2101, 4050, 4055
2010, 4000, 4010
mōurum -um -ūm other endings
V1.5 mōurum:
4000, 4010,
4025, 4040,
4050, 4055
pasūm -um -ūm other endings
V13.31 pasūm: 4000, 4010
V13.32a pasūm: 4010 pasīm: 4000
V13.32b pasūm: 4000, 4010
V13.33a pasūm: 4000, 4010
V13.33b pasūm: 4000, 4010
V13.34 pasūm: 4000, 4010
24
Corresponds to VrS32.5.
25
Corrected into gāθīm.
Book 1.indb 111 24/05/16 13:44

112 a. cantera
pitūm -um -ūm other endings

Y9.11 paitūm: 4010, 4025 pitīm: 15, 82, 83, 2010, 2101
paitīm: 20, 2040, 2104, 4000,
4020, 4031, 4040, 4045, 4050,
4055, 4060, 4070, 4100, 4140,
5020
pərətūm -um -ūm other endings
Y46.10 pərətūm: 2010, 4010 pərətīm: 20, 2101, 4000, 4050,
4055
V13.3 ºpərətūm: 4010 ºpərətīm: 4000, 4050, 4055
V18.6 ºpərətīm: 4000, 4010, 4050,
4055
V19.29 ºpərətūm: 4000, ºpərətīm: 4050, 4055
4010
V19.30 pərətīm: 4000, 4010, 4050,
4055
V19.36 ºpərətūm: 4000 ºpərətīm: 4010, 4050, 4055
VrS8.126 ºpərətūm: 2010, pərətīm: 2101, 4000, 4025,
4010, 4040 4050, 4055, 4060, 4100
Y19.6 pərətūmcit: 2010, pərətīmcit: 20, 2101, 2230,
4025, 4040 4000, 4010, 4050, 4055, 4060,
4100
Y71.16 pərətəm: 20
pərətīm: 2010, 4000, 4010,
4050
rašnūm -um -ūm other endings
VrS8.227 [raš]nūm: 2005 rašnīm: 2101, 2230
rašnūm: 2010, 4040 rašnīm: 4001, 4010, 4050,
rašnūm: 4025 4060, 4100
[raš]nīm: 4055
Y2.7 rašnūm: 2010 rašnīm: 2101, 2104
rašnīm: 20, 2102
Y6.628 rašnūm: 2010 rašnīm: 20, 2101, 2102
rašnīm: 2104
Y16.5 rašnūm: 2010, 4040 rašnīm: 20, 4000, 4010, 4050,
rašnūm: 4025 4055, 4060, 4100
rašnīm: 2101
Y17.629 rašnūm: 2010, 4000, rašnīm: 20, 4010, 4025, 4050,
4040 4055, 4060, 4100
26
Vr7.1
27
Vr7.2
28
= VS6.2
29
= VS17.2
Book 1.indb 112 24/05/16 13:44

ratūm -um -ūm other endings

Y2.2 ratūm: 2010, 4000, ratīm: 20, 2101, 2102, 2104,
4010, 4025, 4040 4050, 4055, 4060, 4100, 4140
Y2.4a ratūm: 2010, 4000, ratīm: 20, 2101, 2102, 2104,
4010, 4025, 4040 4050, 4055, 4060, 4100, 4140
Y2.4b ratūm: 2010, 4000, ratīm: 20, 2101, 2102, 2104,
4010, 4025, 4040 4050, 4055, 4060, 4100, 4140
Y2.5a ratūm: 2010, 4000, ratīm: 20, 2101, 2102, 2104,
4010, 4025, 4040 4050, 4055, 4060, 4100, 4140
Y2.5b ratūm: 2010, 4000, ratīm: 20, 2101, 2102, 2104,
4010, 4025, 4040 4050, 4055, 4060, 4100, 4140
Y9.1 ratūm: 40, 4010, ratīm: 15, 20, 82, 83, 2010,
4025, 4055 2040, 2101, 2104, 2801, 4000,
4031, 4040, 4045, 4050, 4060,
4070, 4100, 4140, 5020
ratī: 4020
Y31.2 ratūm: 2005, 2010, ratīm: 20, 2101, 4050, 4055
4000, 4010, 4025,
4040
V15.9a ºratūm: 4000, 4010 ºratīm: 4050, 4055
V15.9b ºratūm: 4000, 4010 ºratīm: 4050, 4055
tāiium -um -ūm other endings
Y9.21 tāiium: 15,
20, 82, 83,
2010, 2040,
2101, 2104,
4000, 4010,
4020, 4025,
4040, 4045,
4050, 4055,
4060, 4070,
4100, 4140,
5020
Y11.3 tāiiūm: 20, 2010, taiiiīm: 2101, 2104, 4040,
4000, 4010, 4025 4050, 4055, 4060, 4100, 4140
Book 1.indb 113 24/05/16 13:44

114 a. cantera
xratūm -um -ūm other endings

Y28.1 xratūm: 2010, 4010, xratīm: 20, 4050, 4055
4025, 4040
xrətūm: 4000
Y32.9 xratūm: 2005, 2010, xratīm: 20, 2101, 4025, 4050,
4000, 4010, 4040 4055
V18.6 xratūm: 4000 xratīm: 4010, 4050, 4055
VrS24.230 xratūm: 2005, 2010, xratīm: 2101, 2230, 4050, 4055
4000, 4010
Y25.6 xratīm: 20
Y62.4 xratīm: 20
zaṇtūm -um -ūm other endings
Y62.531 zaṇtūmca: 2010 zaṇtīmca: 20, 2005, 2101, 400,
4010, 4050, 4055
Y68.5 zaṇtūmca: 2000, zaṇtīmca: 20, 2101, 4000,
2010 4010, 4050, 4055
V1.15 °zaṇtūm: 4010, ºzaṉtīm: 4000, 4050, 4055
4025, 4040
V8.103 zaṉtīmca: 4000, 4010, 402532,
4050, 4055
V18.52 ºzaṉtīm: 4000, 4010, 4050,
4055
žnūm -um -ūm other endings
V9.22 žnīm: 4000, 4010, 4050, 4055
V9.22 žnīm: 4000, 4010, 4050, 4055
V9.22 žnīm: 4000, 4010, 4050, 4055
V9.23 žnīm: 4000, 4010, 4050, 4055
30
Vr19.1
31
= VrS32.5 in the Wisperad, Widēwdād and Wīštāsp Yašt ceremonies.
32
This is a modern part completing a destroyed half page of the original manuscript.
Book 1.indb 114 24/05/16 13:44

Table 1b Indian Sādes

Y9.19 ahūm: 100, 110, 120, 230, ahū.ma: 231
234, 4200, 4210, 4220,
4240, 4250, 4340, 4360,
4400, 4410, 4420, 4440,
4450, 4500, 4503, 4504,
4506, 4507, 4510, 4515
Y11.10 ahūm: 100, 120, 230, 231,
234, 235, 4210, 4240, 4420,
4400, 4510
Y27.4 ahūm: 100, 120, 230, 234,
235, 4420, 4410
dāhūm33: 234, 4400
Y30.6 ahūm: 100, 120, 230, 231, ahū34: 235, 4410
234, 4200, 4210, 4240,
4420, 4400, 4510, 4515
V2.22a ahūm: 4200, 4210, 4230,
4240, 4420, 4400, 4410,
4510, 4515
V2.22b ahūm: 4200, 4210, 4230,
4240, 4420, 4400, 4410,
4510, 4515
V18.16 ahūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
V8.40 barəšnūm: 4200, 4210,
4240, 4420, 4510, 4515
barəsnūm: 4230, 4410
baraišnūm: 4400
V8.41 barəšnūm: 4200, 4210,
4230, 4240, 4420, 4510,
4515
baršnūm: 4400
barəsnūm: 4410
V9.15 barəšnūm: 4200, 4420,
4510, 4515
barəsnūm: 4210, 4230,
4240, 4410
baršnūm: 4400
33
For dā ahūm.
Before marətānō.
34
Book 1.indb 115 24/05/16 13:44

116 a. cantera

Y62.535 daxīūmca: 230
Y68.5 daxiiumca: daxiiūmca: 4210, 4420, daxvauumca: 100
234, 235, 4400, 4515 dāxiiaomca: 4410
4200, 4510 daxaiiūmca: 4240
daxuiiumca: daxaiiaūmca: 231
120
V8.103 daxiiumca: daxiiūmca: 4420, 4400, dāxiiaomca: 4410
4200, 4210, 4410
4230, 4240,
4515
V18.52 ºdaxiiūm: 4200, 4210, 4230, dāxiiaom: 4410
4240, 4400, 4510
gaōdaiiūm -um -ūm other endings
Vr2.11 gaoδāiiūm: 4200 gaoδāīm: 4230
gaoδāiiūm: 4250, 4420, gaoδāəīm: 4240
4410, 4515 guδāīm: 4400
gaodāiiūm: 4510
Y62.5 ºgātūm: 230
V9.33 ºgātūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
V9.34 ºgātūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4515
ºγātūm: 4510
V15.43 gātum: 4240 gātūm: 4200, 4210, 4230,
4420, 4410, 4400, 4510,
4515
35
= VrS32.5 in the Wisperad, Widēwdād and Wīštāsp Yašt ceremonies.
Book 1.indb 116 24/05/16 13:44


Y18.8 mainiiūm: 4510 ºmainiiū: 120, 234,
235, 4200, 4210, 4230,
4240, 4250, 4420,
4410, 4515
ºmaeniiūn: 230
ºmainiiūn: 100
ºmainaiiū: 231
mainū: 4400
Y47.7 mainiiūm: 4400 ºmainiiū: 230, 100,
120, 235, 234, 4200,
4210, 4420, 4410,
4510, 4515
ºminaiiū: 231
V10.5 mainiiūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
Vr2.8 ºmainiiūm: 4200, 4210, ºmainiiū: 4240, 4410
4230, 4250, 4420, 4410, ºmainiiūš: 4400
4510 ºmainiiəuš: 4515
Vr19.0 ºmainiiūm: 4200, 4230, ºmainiiū: 4240, 4410,
4400, 4510
ºmainiiəuš: 4515
V13.31 pasūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
V13.32a pasūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
V13.32b pasūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
V13.33a pasūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
V13.33b pasūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
V13.34 pasūm: 4200, 4210, 4230,
4240, 4420, 4410, 4400,
4510, 4515
Book 1.indb 117 24/05/16 13:44

118 a. cantera

Y9.11 pitūm: 100, 110, 4360, 4503, paitəm:
4506 4210,4240,4250
patūm: 230, 231, 4340, patəm: 4200
4400, 4420, 4440, 4450, patīm: 4504
4500, 4507, 4510, 4515
paitūm: 120, 234, 4220, 4410
Y46.10 pərətūm: 100, 120, 230, 231,
234, 235, 4200, 4210, 4240,
4410, 4420, 4510, 4515
pairitūm: 4400
V13.3 ºpərətūm: 4200, 4210, 4230,
4240, 4410, 4420, 4510,
4515
ºpairətūm: 4400
V18.6 ºpərətūm: 4200, 4210, 4230,
4240, 4400, 4410, 4420,
4510, 4515
V19.29 ºpərətūm: 4200, 4210, 4230,
4240, 4410, 4420, 4510,
4515
ºpairitūm: 4400
V19.30 pərətūm: 4200, 4210, 4230,
4240, 4400, 4410, 4420,
4510, 4515
V19.36 ºpərətūm: 4200, 4210, 4230,
4240, 4400, 4410, 4420,
4510, 4515
VrS8.1 pərətūm: 4200, 4210, 4230,
4240, 4250, 4400, 4410,
4420, 4510, 4515
Y19.6 pərətūm.cit: 100, 120, 230, pəratuum.cit: 231
234, 4200, 4250, 4410
pərətūm.cīt: 4420. 4515
pərətūmcit: 4210, 4240
pərətūmcīt: 4510
pairi.tūm.cit: 235
pairi.tūm.cīt: 4400
Y71.16 pərətūm: 100, 120, 230, 235,
4200, 4240, 4410, 4420,
4510, 4515
pəritūm: 231
pairitūm: 234, 4400
Book 1.indb 118 24/05/16 13:44


VrS8.2 rašnūm: 4200, 4400, 4420
rasnūm: 4210, 4230, 4240,
4250, 4410, 4510, 4515
Y2.7 rasnūm: 100, 120, 230, 234,
235
rašnūm: 231, 4230
Y16.5 rasnūm: 100, 230, 235,
4200, 4210, 4240, 4250,
4410, 4515
rašnūm: 120, 234, 4400,
4420, 4510
rašanūm: 231
Y2.2 ratūm: 100, 120, 231, 234,
235, 4200, 4230, 4240, 4250,
4420, 4410, 4510, 4515
rtūm: 4400
Y2.4a ratūm: 100, 120, 231, 234,
235, 4200, 4240, 4250,
4420, 4410, 4510, 4515
rtūm: 4400
Y2.4b ratūm: 100, 120, 231, 234,
235, 4200, 4240, 4250,
4420, 4410, 4510, 4515
rtūm: 4400
Y2.5a ratūm: 100, 120, 231, 234,
235, 4200, 4240, 4250,
4420, 4410, 4510, 4515
rtūm: 4400
Y2.5b ratūm: 100, 120, 231, 234,
235, 4200, 4240, 4250,
4420, 4410, 4510, 4515
rtūm: 4400
Y9.1 ratūm: 100, 110, 120, 230,
231, 234, 235, 4200, 4210,
4220, 4240, 4250, 4340,
4360, 4400, 4410, 4420,
4440, 4450, 4500, 4503,
4504, 4506, 4507, 4510, 4515
Y31.2 ratūm: 120, 231, 234, 235, ratuum: 100
4200, 4210, 4240, 4410,
4420, 54510, 4520
rtūm: 4400
Book 1.indb 119 24/05/16 13:44

120 a. cantera
V15.9a ratūm: 4200, 4210, 420,

4240, 4410, 4420, 4510,
4520
rtūm: 4400
V15.9b ratūm: 4200, 4210, 420,
4240, 4410, 4420, 4510,
4520
rtūm: 4400
Y9.21 tāiiūm: 110, 120, 234, 235, tāīuum: 100, 230
4200, 4210, 4220, 4240, tāiim: 4400
4250, 4340, 4360, 4410, tāiiū: 231
4420, 4440, 4450, 4500,
4503, 4504, 4506, 4507,
4510, 4515
Y11.3 tāiiūm: 230, 100, 120, 234, tāiiū: 231
235, 4200, 4210, 4240,
4250, 4420, 4410, 4400,
4510, 4515
xratūm -um -ūm other endings
Y28.1 xratūm: 100, 120, 234, 235, xratū: 230
4200, 4210, 4240, 4250,
4410, 4420, 4510, 4515
xaritūm: 231
xrətūm: 4400
Y32.9 xratūm: 100, 120, 230, 234, xaratū: 4510
235, 4200, 4210, 4240, 4420 xratū: 4515
xritūm: 231
xaratūm: 4410
xarətūm: 4400
V18.6 xratūm: 4200, 4210, 4230,
4240, 4400, 4420, 4510,
4515
xaratūm: 4410
VrS24.2 xratūm: 4200, 4230, 4240,
4510, 4515
xaratūm: 4410
xritūm: 4400
Y25.6 xratūm: 100, 120, 230, 234
xritūm: 231
xaratūm: 235
Book 1.indb 120 24/05/16 13:44

Y62.4 xrətūm: 230

Y62.5 jṇtūmca: 230
Y68.5 zaṇtūmca: 230, 100, 120,
231, 234, 235, 4200, 4210,
4240, 4420, 4410, 4400,
4510, 4515
V1.15 ºzaṇtūm: 4200, 4210, 4230, ºjaṇtū: 4400
4240, 4250, 4420, 4510,
4515
ºjaṇtūm: 4410
V8.103 zaṇtumca: zaṇtūmca: 4200, 4210, 4230, zaṇtąmca: 4420
4515 4240, 4400, 4410
V18.52 ºzaṇtūm: 4200, 4210, 4230,
4240, 4400, 4510
V9.22 žnūm: 4200, 4210, 4230,
4240, 4420
žanūm: 4400, 4410, 4510,
4515
V9.22 žnūm: 4200, 4210, 4230,
4240, 4420
žanūm: 4400, 4410, 4510,
4515
V9.22 žnūm: 4200, 4210, 4230,
4240, 4420
žanūm: 4400, 4410, 4510,
4515
V9.23 žnūm: 4200, 4210, 4230,
4240, 4420
žanūm: 4400, 4410, 4510,
4515
Book 1.indb 121 24/05/16 13:44

122 a. cantera
Table 1c Combined
Yasna
Y9.19 ahūm: 400, 408, 409, 410,
415, 420, 450
Y11.10 ahūm: 400, 415, 420
Y27.4 ahūm: 400, 415, 420
Y30.6 ahūm: 400, 415, 420
Y62.5 daxiiūmca: 400, 415, 420
Y68.5 daxiiūmca: 400, 415, 420
Y62.5 ºgātūm: 400, 415, 420
maińiiūm -um -ūm other endings
Y18.8 ºmainiiū 400
ºmainiiū: 415, 420
Y9.11 patūm: 408, 409, 420 paitīm: 400
paitūm: 410 paitəm: 415
patīm: 450
Y46.10 pərətūm: 420
fərətūm: 400, 415
Y19.6 pərətūm.cit: 400
pairitūm.cīt: 415
pərətəmcīt: 420
Y71.16 pərətūm: 400, 415, 420
Y2.7 rašnūm: 400, 415, 420
Y6.6 rašnūm: 400
rasnūm: 415
Y16.5 rašnūm: 400, 415, 420
Y17.6 rašnūm: 400
Y2.2 ratūm: 400, 415, 420
Y2.4a ratūm: 400, 415, 420
Y2.4b ratūm: 400, 415, 420
Y2.5a ratūm: 400, 415, 420
Book 1.indb 122 24/05/16 13:44

Y2.5b ratūm: 400, 415, 420

Y9.1 ratūm: 400, 408, 409, 410,
415, 420, 450
Y31.2 ratūm: 400, 415, 420
Y9.21 tāiium: 408, tāiiūm: 400
409, 410, 415,
420, 450
Y11.3 tāiiūm: 400, 415, 420
xratūm -um -ūm (/-īm) other endings
Y28.1 xratūm: 400, 415, 420
Y32.9 xratūm: 400, 415, 420
Y25.6 xratūm: 400, 420
Y62.4 xratūm: 400, 415
xrāitūm: 420
Y62.5 zaṇtumca: 415 zaṇtūmca: 400, 420
Y68.5 zaṇtūmca: 400, 415, 420
Table 1d Yasna Pahlavi

Y9.19 ahūm: 500, 602, 605, 613 ahū: 510, 530
Y11.10 ahūm: 500, 510, 530, 613
Y27.4 ahūm: 500, 510, 530
Y30.6 ahūm: 500, 510, 530
Y62.5 daxiiūmca: 500, 510, 530
Y68.5 daxiiūmca: 500, 510, 530
Y62.5 ºgātūm: 510, 530 ºgātū: 500
mańiiūm -ūm
Y18.8 mainiiūm: 500, 510, 530
Y9.11 patūm: 510, 530, 602 paitīm: 613
paitūm: 605
Book 1.indb 123 24/05/16 13:44

124 a. cantera

Y46.10 pərətūm: 500, 510, 530
Y19.6 pərətūmcīt: 500, 510
pərətūm.cit: 530
Y71.16 pərətūm: 500, 510, 530
raṣ̌nūm -um -ūm other endings
Y2.7 rašnūm: 500, 510
Y6.6 rašnūm: 530, 510
Y16.5 rašnūm: 500, 510, 530
Y2.2 ratūm: 500, 510, 530
Y2.4a ratūm: 500, 510
Y2.4b ratūm: 500, 510
Y2.5a ratūm: 500, 510
Y2.5b ratūm: 500, 510
Y9.1 ratūm: 500, 510, 530, 602,
605, 613
Y31.2 ratūm: 500, 510, 530
Y9.21 tāiium: 500, tāiiūm: 510, 530, 605, 613
602
Y11.3 tāiiūm: 500, 510, 530, 613
Y28.1 xratūm: 500, 510, 530
Y32.9 xratūm: 500, 510, 530
Y25.6 xratūm: 500, 510, 530
Y62.4 xratūm: 500, 510, 530
zaṇtūm
Y62.5 zaṇtūmca: 500, 510, 530
Y68.5 zaṇtūmca: 500, 510, 530
Book 1.indb 124 24/05/16 13:44

Table 1e Yasna Sanskrit

Y9.19 ahūm: 677, 682
Y11.10 ahūm: 677, 682
Y27.4 ahūm: 680 ahū: 682
Y30.6 ahūm: 677, 680, 682
Y62.5 daxiiumca: daxiiūmca: 680
682
Y68.5 daxiiūmca: 682
Y62.5 ºgātūm: 680, 682
Y18.8 mainiiūm: 677 mainiiū: 682
Y46.10 pərətūm: 677, 680, 682
Y19.6 pərətūmcit: 677
Y71.16 pərətūm: 682
Y9.11 pətūm: 677 patīm: 682
raṣ̌nūm -um -ūm other endings
Y2.7 rasnūm: 677, 682
rašnūm: 681
Y6.6 rasnūm: 677
rašnūm: 681
Y16.5 rašnūm: 677, 682
Y2.2 ratūm: 677, 681 ratuum: 682
Y2.4a ratūm: 677, 681, 682
Y2.4b ratūm: 677, 681, 682
Y2.5a ratūm: 677, 681, 682
Y2.5b ratūm: 677, 681, 682
Y9.1 ratūm: 677,681,682
Y31.2 ratūm: 677, 680, 682
Y9.21 tāiium: 677 tāiiūm: 682
Y11.3 tāiiūm: 682
Book 1.indb 125 24/05/16 13:44

126 a. cantera

Y28.1 xratūm: 677, 680, 682
Y32.9 xratūm: 677, 680, 682
Y25.6 xratūm: 682
Y62.4 xratūm: 680, 682
Y62.5 zaṇtūmca: 680, 682
Y68.5 zaṇtūmca: 682
Table 1f Wīdēwdād
Pahlavi
V2.22a ahūm: 4670
V2.22b ahūm: 4670, 4700, 4711,
4712, 4713, 4715, 4710
V18.16 ahūm: 4600, 4660, 4700,
4711, 4712, 4713, 4715,
4710
V8.40 barəšnūm: 4601 barəšnū.ąm: 4700
baršnūm: 4713 barəšnąm: 4610
barəšnūm: 4670, 4711
baraišnūm: 4700
V8.41 barəšnūm: 4713
barəšnūm: 4700, 4711, 4610
baršnūm: 4601
baršnūm: 4670
V9.15 baršnūm: 4600, 461036,
4700, 4715
barašnūm: 4670, 4711, 4710
baršnūm: 4713
V8.10337 daxiiumca: daxiiūmca: 4700, 4700,
4601 4711, 4713
36
4610 .
37
4610
Book 1.indb 126 24/05/16 13:44

V18.52 ºdaxiium: ºdaxīūm: 4660

4670, 4700,
4712, 4713
V9.34 ºgātūm: 4600, 4670, 4700,
4711, 4713, 4715, 4710
V15.43 gātum: 4600, gātūm: 4670, 4660, 4713 gātąm: 4610, 4711,
4700, 4710 4715
V10.5 mainiiūm: 4600, 4670, 4700,
4711, 4713, 4715, 4710;
m[ainii]ūm: 4610
V13.31 pasūm: 4600, 4610, 4670,
4660, 4700, 4711, 4713,
4715, 4710
V13.32a pasum: 4600, pasūm: 4670, 4713, 4715
4610, 4660,
4700, 4711,
4710
V13.32b pasum: 4600, pasūm: 4610, 4670, 4711,
4660, 4700, 4713, 4715
4710
V13.33a pasum: 4700 pasūm: 4600, 4610, 4670,
4660, 4711, 4713, 4715,
4710
V13.33b pasum: 4700 pasūm: 4600, 4610, 4670,
4660, 4711, 4713, 4715
V13.34 pasum: 4610, pasūm: 4600, 4670, 4660,
4711, 4715 4700, 4713, 4710
V13.3 ºpərətūm: 4600, 4610, 4660,
4670, 4700, 4710, 4711,
4713, 4715
V18.6 ºpərətūm: 4600, 4660, 4670,
4700, 4710, 4711, 4712,
4715
V19.29 ºpərətūm: 4660, 4670, 4700, ºpərətąm: 4600, 4610,
4712, 4713 4710, 4711, 4715
Book 1.indb 127 24/05/16 13:44

128 a. cantera
V19.30 pərətūm: 4600, 4610, 4660,

4670, 4700, 4710, 4711,
4712, 4713, 4715
V19.36 ºpərətūm: 4600, 4610, 4660,
4670, 4700, 4710, 4711,
4712, 4713, 4715
V15.9a ratum: 4600, ratūm: 4660, 4670, 4700,
4610, 4711, 4710
4713, 4715
V15.9b ratum: 4711, ratūm: 4600, 4660, 4670,
4713, 4715 4700, 4710
V18.6 xratūm: 4600, 4610, 4660,
4670, 4700, 4710, 4711,
4712, 4715
V1.15 ºzaṇtum: 4713, ºzaṇtūm: 4600, 4670, 4700, ºzaṇtəm: 4601
4700 4711, 4712, 4715, 4710
V8.10338 zaṇtumca: zaṇtūmca: 4700, 4711
4601, 4713
V18.52 °zaṇtūm: 4660, 4670, 4700,
4712, 4713
V9.22 žnūm: 4600, 4670, 4700;
[žn]ūm: 4610
žanūm: 4713
V9.22 žnūm: 4600, 4610, 4670,

4700, 4713
38
4610 .
Book 1.indb 128 24/05/16 13:44

V9.22 žnūm: 4600, 4610, 4670,

4700, 4710, 4711, 4715
žanūm: 4713
V9.23 žnūm: 4600, 4610, 4670,
4700, 4710, 4711, 4715
žanūm: 4713
Table 2 : ū- stems
fsəratūm -um -ūm other endings
Y5.5 fsəratūm: 4010, 4025, 4040 fsəratīm: 4060, 4100,
fsiratūm: 4000 4140
fsarətūm: 2010 fsarətīm: 2101
fsiratīm: 2104, 4050,
4055
Y27.9 fsəratūm: 2005, 4000, 4010, fsəratīm: 2101, 4025,
4040 4050, 4100
fsaratūm: 2010 pəsəratīm: 20
fsiratūuuōm: 4055
Y33.12 fsəratūm: 4040
Y37.5 fsəratūm: 2005, 4000, 4010, fsəratīm: 2101,
4025
tanūm -um -ūm other endings
Y9.17 ºtanīm: 15, 20, 82, 83,
2010, 2040, 2101,
2104, 2801, 4000,
4010, 4020, 4025,
4031, 4040, 4050,
4055, 4060, 4070,
4100, 4140, 5020
ºtanəm: 4045
Y10.14 tanūm: 2010, 4000, 4010, taniīm: 20, 2101, 2104,

4025, 4040 4050, 4055, 4060,
4100, 4140
Book 1.indb 129 24/05/16 13:44

130 a. cantera
Y33.10 tanūm: 40, 2005, 2010, tanīm: 2101, 4050,

4000, 4010, 4025, 4040 4055
V4.17 tanūm: 4000, 4010, 4025, tanīm: 4050, 4055

4040
V5.12 tanūm: 4000, 4010, 4040 tanīm: 4025, 4050,
4055
V9.1 tanūm: 4000 tanīm: 4010, 4050,
4055
xšnūm -um -ūm other endings
Y48.12 xšnūm: 2005, 4000, 4010 xšnīm: 20
xšnīm: 2101
xšnəm: 2010, 4000,
4010, 4050, 4055
Y53.2 xšnūm: 2005, 2010 xšnīm: 20, 4050, 4055

xšnəm: 2101

Y5.5 fsəratūm: 230, 100, 4200,
4210, 4240, 4250, 4420
fšəratūm: 120, 234, 235
fšə.ratūm: 4510
fšə.ratūm: 4400, 4515
frsəratūm: 100
frašratūm: 231
frasaratūm: 4410
Y27.9 fširatūm: 230
frširatūm: 100
fərəšaratūm: 231
fəsəratūm: 235
fəsə.ratūm: 4240
Y37.5 fəsirətūm: 230
fsarətūm: 100
Book 1.indb 130 24/05/16 13:44


Y9.17 ºtanūm: 100, 110, 120, 230, ºtəməm: 4250
231, 234, 235, 4200, 4210,
4220, 4240, 4340, 4400,
4420, 4440, 4450, 4503,
4504, 4506, 4507, 4510,
4515
Y10.14 tanūm: 230, 100, 120, 231,
234, 235, 4200, 4210, 4240,
4250, 4420, 4410, 4400,
4510, 4515
Y33.10 tanūm: 235, 234, 4200, tanōm: 100, 230, 231
4410, 4400, 4510, 4515 tanəm: 120
tūm: 4240
V4.17 tanūm: 4200, 4210, 4230,
4240, 4250, 4420, 4410,
4400, 4510, 4515
V5.12 tanūm: 4200, 4210, 4230,

4240, 4420, 4410, 4400,
4510, 4515
V9.1 tanūm: 4200, 4210, 4230,

4240, 4420, 4410, 4400,
4510, 4515
Y48.12 xšnūm: 231, 4400 xšnəm: 120, 234, 235,
4200, 4210, 4240
xšnəm: 4420, 4510,
4515
xšnəm: 230, 100
xašnəm: 4410
Y53.2 xšnəm: 234, 235, 4210,

4240, 4420
xšnəm: 230, 100, 120
xšnəm: 4400, 4510,
4515
xašnəm: 4410
xašnēm: 231
Book 1.indb 131 24/05/16 13:44

132 a. cantera
Table 2c Combined
Yasna
Y5.5 fəšəratūm: 415 fsəratīm: 400
fsə.ratūm: 420
Y27.9
Y9.17 ºtanūm: 400, 408, 409, 410,
415, 420, 450
Y10.14 tanūm: 400, 415, 420
Y33.10 tanūm: 400, 415, 420
Y48.12 xšnəm: 400, 420
xšnəm: 415
Y53.2 xšnəm: 400
xšnəm: 415, 420

Y5.5 fsəratūm: 500, 510, 530
fəsarəatūm: 510, 530
Y37.5 fsəratūm: 500, 510, 530
Y9.17 ºtanūm: 500, 510, 530, 602,
605, 613
Y10.14 tanūm: 500, 530, 613
Y33.10 tanūm: 510, 530
Y48.12 xšnūm: 510, 530 xšnəm: 500
Y53.2 xšnūm: 500, 530 xšnəm: 510
Book 1.indb 132 24/05/16 13:44


fəšəratūm: 681
fsərətūm: 682
Y27.9 fšəratūm: 677, 682
fsə.ratūm: 680
Y33.12 fsəratūm: 677, 682
Y9.17 ºtanūm: 677, 682
Y10.14 tanūm: 677, 682
Y33.10 tanūm: 677, 680, 682
Y48.12 xšnūm: 680 xšnəm: 682
Y53.2 xšnəm: 680
xšnəm: 682
Pahlavi
V4.17 tanūm: 4600, 4670, 4700,
4711, 4712, 4713, 4715,
4710
V5.12 tanūm: 4601, 4670, 4700,
4711, 4712, 4713, 4715,
4710
V9.1 tanūm: 4600, 4670, 4700,
4711, 4713, 4715, 4710
Book 1.indb 133 24/05/16 13:44

134 a. cantera
Table 3: ṷa- stems

caθrušum -um -ūm other endings
Y19.7 caθrušum: 2005, 4000, 4055, caθrušąm: 2010, 4040
4060, 4100
ciθrāšum: 20
ciθrušum: 2101, 4025, 4050
V6.32 ciθrušum: 4000, 4055 caθrušąm: 4010, 4025, 4040,
4601
ciθrušąm: 4050
V16.2 caθrušum: 4010, 4050 caθrušąm: 4000
ciθrušum: 4055
pourum -um -ūm other endings
V4.47 pourum: 4000, 4010, 4025,
4040, 4055
paōurum: 4050
V8.40 pourum: 4000, 4055, 4200
pōurum: 4010, 4050
V8.41 pourum: 4000, 4010, 4025, 4055
pōurum: 4050
V8.58a pourum: 4000, 4010, 4025, 4055
pōurum: 4050
V8.58b pourum: 4000, 4010, 4025
pōurum: 4050, 4055
V9.15 paōurum: 4000, 4010, 4055 paōirīm: 4050
saurum -um -ūm other endings
V10.9 saōrəm: 4000, 4010, 4050
sōrəm: 4055
θrišum -um -ūm other endings
Y19.7 θrišum: 4000, 4040, 4050, 4055 θrišąm: 2010, 4025
θrəšum: 20, 2005, 2101, 4100
srišum: 4060
V6.32 θrišum: 4000, 4010, 4025, 4040, θrišąm: 4050
4055
V16.2 θrišąm: 4000, 4010, 4050,
4055
V18.63a θrušąm: 4000, 4010, 4055
θrišąm: 4050
Book 1.indb 134 24/05/16 13:44

V18.63b θrišum: 4000, 4050

θrəšum: 4010, 4055
V18.64a θrəšum: 4010, 4055 θrišąm: 4000
θrišum: 4050
V18.64b θrišum: 4000, 4050
θrəšum: 4010, 4050

Y19.7 ciθrušum: ciθrušūm: 234, 4240, 4420, ciθarašuum: 231
230, 100 4410 ciθrušəm: 4510, 4515
caθrušūm: 235, 4200, 4210,
4350
ciθrašūm: 4400
V6.32 caθrušūm: 4200, 4210, ciθrušəm: 4510, 4515
4230, 4240, 4420, 4400
ciθrušūm: 4410
V16.2 caθrušum: caθrušūm: 4230, 4420 ciθrušąm: 4400
4200, 4210 ciθrušūm: 4410 ciθrušəm: 4510, 4515
ciθrušum:
4240
haurum -um -ūm -əm
Y19.14 ahurəm: 100, 120, 230,
231, 234, 235, 4200,
4210, 4240, 4250,
4410, 4510, 4515
hurəm: 4400
haurəm: 4420
Y20.3 ahurəm: 100, 120, 230,
231, 234, 235, 4200,
4210, 4250, 4400,
4410, 4510, 4515
V13.12 ahurum: 4240 haurəm: 4420
V13.13 °haourūm: 4200, 4210,
4240, 4410, 4510, 4515
°hurūm: 4400
°haorūm: 4410
Book 1.indb 135 24/05/16 13:44

136 a. cantera
V13.20 °haurūm: 4200, 4230, 4420,

4515
°haorūm: 4210, 4410
°harūm: 4240
°huruum: 4400
°haourūm: 4510

V4.47 pōurum: 4200,
4210, 4230,
4250, 4420,
4410, 4515
paourum:
4240, 4400
paōurum:
4510
V8.40 paourum: paourūm: 4400 paoirīm: 4210, 4410

4420, 4510, paōirīm: 4240
4515
pourum: 4200
paorum: 4230
V8.41 pourum: 4200 paourəm: 4230

paourum: paoirīm: 4210, 4410
4420, 4400, paōirīm: 4240
4510, 4515
V8.58a pourum: 4200 paoirīm: 4230
pōurum: 4240,
4420
paōurum:
4210
paourum:
4410, 4400,
4510
purum: 4515
V8.58b pourum: 4200 paourəm: 4230
pōurum: 4210,
4240, 4420,
4515
paourum:
4400, 4510
Book 1.indb 136 24/05/16 13:44

V9.15 paōurum: paōirīm: 4210, 4230,

4200 4240
paourum: paoirīm: 4410
4420, 4400,
4510, 4515

V10.9 saurum: 4420 saurəm: 4200
saorum: 4210 surəm: 4230, 4510,
surum: 4410, 4515
4400
Y19.7 θrišūm: 120, 234, 4200, θrəšuum: 230
4420, 4510, 4515
θrəšūm: 100, 235, 4210,
4240, 4250
θarišūm: 231
θaraešūm: 4410
θrašūm: 4400
V6.32 θrišum: 4200, θrišūm: 4230, 4240, 4410,

4210 4420, 4400, 4510, 4515
θrəšūm: 4710
V16.2 θrišūm: 4200, 4210, 4230,
4240, 4420, 4400, 4510,
4515, 4670
θarišūm: 4410
V18.63a θrišūm: 4200, 4230, 4420,
4410, 4510, 4515
θrišūm: 4210, 4240
V18.63b θrišūm: 4200, 4230, 4420,
4400, 4510, 4515
θarišūm: 4410
θrišūm: 4210, 4240
V18.64a θrišūm: 4200, 4230, 4420,
4510, 4515
θarišūm: 4410
θraišūm: 4400
θrišūm: 4210, 4240
V18.64b θrišūm: 4200, 4230, 4420,

4410, 4400, 4510, 4515
θrišūm: 4210, 4240
Book 1.indb 137 24/05/16 13:44

138 a. cantera
Table 3c Combined
Yasna
caθrušum -um -ūm -əm
Y19.7 ciθrušum: 400 ciθrašūm: 420
Y19.14 haurum: 400 ahurəm: 415
haurəm: 420
Y20.3 ahurəm: 415
θrišum -um -ūm -əm
Y19.7 θrišūm: 400, 415, 420
Table 3d Pahlavi Yasna

Y19.7 [ca]θrušūm: 500 caθrušəm: 510, 530
Y19.14 haurəm: 500, 510, 530
Y20.3 haurum: 510, haurəm: 500
530
Y19.7 θrišum: 500 θrišūm: 510, 530

Y19.7 caθrušum:
677
ciθrušum: 680
ciθrušum: 682
haurum -um -ūm other endings
Y19.14 ahurəm: 680
ahurim: 682
Y20.3 ahurəm: 680, 682
Y19.7 θrišum: 677,
682
θrišum: 680
Book 1.indb 138 24/05/16 13:44

Table 3f Pahlavi
Wīdēwdād
V6.32 caθrušūm: 4710 caθrušąm: 4601
caθrušąm: 4610, 4711,
4715
ciθrušąm: 4712
ciθrušąm: 4700, 4713
ciθrušəm: 4670
V16.2 caθrušum: caθrušąm: 4600, 4660,
4610, 4711, 47 4710
caθrušəm: 4670
haurum -um -ūm other endings
V13.12 °haurum:
4600, 4610,
4660, 4670,
4700, 4711,
4713
V13.13 °haurum:
4600, 4610,
4660, 4670,
4700, 4711,
4713, 4715
V4.47 paourum: paoru: 4712
4601, 4670,
4700, 4713,
4715, 4710
V8.40 paourum: paourūm: 4601, 4670, 4700
4610, 4711,
4713, 4715
V8.41 paourum:
4601, 4610,
4670, 4700,
4711, 4713,
4715
V8.58a pōurum: 4601,
4670
paourum:
4610, 4700,
4711, 4713,
4715
Book 1.indb 139 24/05/16 13:44

140 a. cantera
V8.58b pōurum: 4601,

4670
paōurum:
4700
paourum:
4610, 4711,
4713, 4715
V9.15 paourum: paoirīm: 4670
4600, 4610,
4700, 4710,
4711, 4713,
4715
V10.9 saurum: 4610; surəm: 4700 sauru:
surum: 4670, 4600
4711, 4715 suru: 4710
V6.32 θrišum: 4610 θrišūm: 4670, 4711 θrišąm: 4700, 4713
θrəšūm: 4710 θrišąm: 4601, 4712
V16.2 θrišum: 4610, θrišūm: 4670 θrišąm: 4700, 4713
4711, 4715 θrəšūm: 4710 θrišąm: 4600, 4660
V18.63a θrišūm: 4660, 4713 θrišąm: 4600, 4610,

θrišūm: 4670 4712, 4715
θrašąm: 4700
V18.63b θrišūm: 4660 θrišąm: 4600, 4700,
θrišūm: 4610, 4670, 4713 4712
θrišīm: 4715
V18.64a θrišum: 4600, θrišūm: 4660, 4713
4700, 4712 θrišūm: 4610, 4670, 4715
V18.64b θrišūm: 4660, 4713 θrišąm: 4600, 4700,
θrišūm: 4610, 4670, 4715 4712
Book 1.indb 140 24/05/16 13:44

Table 4: xštūm, srūm, dādrūm, drūm

dādrūm -um -ūm other endings
V9.11 dādrīm: 4050, 4055
dādrə: 4000
dādrī: 4010
drum -um -ūm other endings
V22.5 drūm: 4000, 4010 drīm: 4050, 4055
srūm -um -ūm other endings
V9.14a srīm: 4000, 4010,
4050, 4055
V9.14b srīm: 4000, 4010,
4050, 4055
V9.14c srīm: 4000, 4010,
4050, 4055
V16.6 sraēm: 4000
srīm: 4050, 4055
xštūm39 -um -ūm other endings
Y9.21 xštūm: 4000, 4010, 4025 xštīm: 15, 20, 82, 83,
2010, 2040, 2101,
2104, 4000, 4020,
4040, 4045, 4050,
4055, 4060, 5020
xstīm: 4070, 4100,
4140
xvštīm: 4031
V4.19 xštūm: 4010 xštīm: 4000, 4025,
4040, 4050, 4055
V4.23 xštūm: 4000, 4010 xštīm: 4025, 4040,
4050, 4055
V5.29 xštīm: 4000, 4010,
4025, 4040, 4050, 4055
39
Similar distribution in V5.30, 8.77, 9.7, 9.28.
Book 1.indb 141 24/05/16 13:44

142 a. cantera

V9.11 dādrūm: 4200, 4400, 4670 dāδrəm: 4410
dāδrūm: 4210, 4230, 4510,
4515
dāδarūm: 4240
dādarūm: 4420
drūm -um -ūm other endings

V22.5 drūm: 4200, 4230, 4420, dūrəm: 4210, 4240
4510, 4515
darūm: 4410
δrūm: 4400
V9.14a srūm: 4200, 4210, 4230,
4240, 4420, 4400, 4510,
4515
sarūm: 4410
V9.14b srūm: 4210, 4230, 4240,
4420, 4510, 4515
sarūm: 4410
šrūm: 4400
V9.14c srūm: 4210, 4230, 4240, sūrəm: 4200

4420, 4400, 4510, 4515
sarūm: 4410
V16.6 srūm: 4200, 4210, 4240,

4420, 4410, 4400, 4510,
4515
sūrūm: 4230
xštūm -um -ūm other endings
Y9.21 xštūm: 100, 110, 120, 230, xštəm: 4450
234, 235, 4200, 4210, 4220, xštū: 235
4240, 4340, 4360, 4400,
4420, 4440, 4500, 4503,
4504, 4506, 4507, 4510,
4515
xstūm: 4250
xaštūm: 4410
xaštūm: 231
Book 1.indb 142 24/05/16 13:44

V4.19 xštūm: 4010, 4210, 4230, xštū: 4200

4240, 4250, 4420, 4400,
4510, 4515
xaštūm: 4410
V4.23 xštūm: 4000, 4010, 4200,
4210, 4230, 4240, 4250,
4420, 4400, 4510, 4515
V5.29 xštūm: 4200, 4210, 4230,
4240, 4250, 4420, 4400,
4510, 4515
Table 4c Combined
Yasna
Y9.21 xštūm: 400, 408, 409, 410,
415, 450
Y9.21 xštūm: 500, 510, 530, 602,
605, 613
Y9.21 xštūm: 677, 682
Pahlavi
V9.11 dādrum: 4601, dādarąm: 461040, 4711,
4700, 4713 4715
dādarum:
4710
drūm -um -ūm other endings
V22.5 drum: 4670, drūm: 4600, 4700, 4710,
4715 4712, 4713
darūm: 4660
40
4610 .
Book 1.indb 143 24/05/16 13:44

144 a. cantera

V9.14a srum: 4600, srūm: 4710
4610, 4670,
4700, 4711,
4713, 4715
V9.14b srum: 4600,
4610, 4670,
4700, 4711,
4713, 4715,
4710
V9.14c srum: 4600,
4610, 4670,
4700, 4711,
4713, 4715,
4710
V16.6 srum: 4600,
4670, 4660,
4700, 4711,
4713, 4715,
4710
V4.19 xštūm: 4601, 4670, 4700,
4711, 4712, 4713, 4715,
4710
V4.23 xštūm: 4601, 4670, 4700,

4711, 4712, 4713, 4715,
4710
V5.29 xštūm: 4601, 4610, 4670,

4700, 4711, 4712, 4713,
4715, 4710
Book 1.indb 144 24/05/16 13:44

Table 5: ai̭ṷa- stems

harōiium -um -ūm other endings
V1.8 harōiium: 4010, 4025, 4040,
4055
hrōiium: 4000
haraōiium: 4050
hōiium -um -ūm other endings
V9.17a hōiium: 4000, 4010, 4055
haōiium: 4050
V9.17b hōiium: 4000, 4010
haōiium: 4050, 4055
V9.18a hōiium: 4000, 4010, 4055 haōiiąm: 4050
V9.18b hōiium: 4000, 4010, 4055 haōiiąm: 4050
hōiiumca
Y11.4 hōiiumca: 20, 2101, 2104, 4000,
4010, 4050
haōiiumca: 2010, 4025, 4040,
4055, 4060, 4100, 4140
Y11.5 hōiiumca: 20, 2101, 4010, 4025, hōiiąmca: 405041
4040, 4050, 4060
haōiiumca: 2010, 2104, 4000,
4055, 4100, 4140
ōiium -um -ūm other endings
V8.25 ōiium: 4000, 4050, 4055
aōiium: 4010, 4025
V16.12 ōiium: 4025, 4050
aōiium 4000, 4010, 4055
V20.4 aōiium: 4000, 4010, 4050, 4055
41
Interestingly, the scribe copied Y11.5 twice, each time spelling hōiiumca differ-
ently: first hōoiiąmca and then hōiiumca.
Book 1.indb 145 24/05/16 13:44

146 a. cantera
vīdōiium -um -ūm other endings

Y2.13 vīdōiium: 20, 2010, 2101, 2102,
2104, 4040, 4050, 4055, 4060,
4100, 4140
vīdoiium: 4000, 4010, 4025
Y6.12 vīdōiium: 20, 2101, 2104, 4010, vīduiiąm: 2010
4025, 4040, 4050, 4055, 4060
vīdoiium: 4000, 4100, 4140
*Y17.13 vīdōiium: 20, 4000, 4010, 4050,

4055, 4060, 4100
vīdium: 4025
*Y25.6 vīdōiium: 20
*Y71.5 vīdōiium: 2005, 2010, 2101,
4000, 4050
vīdaōiium: 4010
vīdīiium: 20
V5.22 vīdōiium: 4000, 4040 vīduiiōm: 4010, 4025
vīduiium: 4050, 4055
V5.23 vīdōiium: 4010, 4025, 4040,
4055
vīduiium: 4050
V5.24a vīdōiium: 4000, 4025, 4040,

4050
vīdaōiium: 4010, 4055
V5.24b vīdōiium: 4000, 4050 vīduiiōm: 4010, 4025, 4040,
4055
V5.25 vīdōiium: 4000, 4010, 4025,
4050, 4055
Book 1.indb 146 24/05/16 13:44


V1.8 harōiiūm: 4200, 4210, 4230,
4240, 4250, 4420, 4410
hurōiiūm: 4400
harōi.yūm: 4510
harō.yūm: 4515
V9.17a hōim: 4200, 4230,
4420, 4410, 4400,
4510, 4515
haōim: 4210, 4240
V9.17b hōim: 4200, 4230,
4240, 4420, 4410,
4400, 4510, 4515
hōm: 4210
V9.18a hōim: 4200, 4210,
4230, 4240, 4420,
4410, 4400, 4510, 4515
V9.18b hōim: 4200, 4210,
4230, 4240, 4420,
4410, 4400, 4510, 4515
hōiiumca
Y11.4 hōiiūmca: 100, 120, 234, hōīmaca: 231
235, 4210, 4240, 4250,
4420, 4410, 4400, 4510,
4515
haōiiūmca: 4200
hūiiūmca: 230
Y11.5 hōiiūmca: 230, 100, 120, hōiimaca: 231
234, 235, 4200, 4210, 4240,
4250, 4420, 4410, 4400,
4510, 4515
V8.25 oim: 4200, 4210,
ōim: 4230, 4240, 4510,
4515
V16.12 ōim: 4200, 4210, 4230,
4400, 4240, 4420
aōim: 4410, 4510, 4515
Book 1.indb 147 24/05/16 13:44

148 a. cantera
V20.4 aōim: 4200, 4210,

4410, 4510
ōim: 4230, 4515
aoim: 4240, 4400
Y2.13 vīdōiiūm: 230, 100, 234, vīduiim: 231
235, 4240, 4250, 4420, vīδuiim: 4400
4510, 4515 vīdaoīm: 4200
vīduiiūm: 120
viδōiiūm: 4410
Y6.12 vīdūīm: 4400
Y25.6 vīduiim: 230, 100
vīdaeuum: 120
vīδuiim: 231
viidaiuum: 235
Y71.5 vīdōiiūm: 234, 4240, 4420, vīdōīm: 230, 120
4510, 4515 βiduiim: 231
vīduiiūm: 100 vīdōiiuum: 235
viδōiiūm: 4200, 4410 vīdūīm: 4400
V5.22 vīdōiiūm: 4200, 4210, 4230,
4420, 4510, 4515
vīδōiiūm: 4410
vīδuiiūm: 4400, 4670
vīduiiūm: 4240, 4400, 4712
V5.23 vīdōiiūm: 4200, 4210, 4230,
4240, 4420, 4510, 4515
vīδuīūm: 4400
vīδōiiūm: 4410
V5.24a vīdōiiūm: 4200, 4210, 4230,
4420, 4510, 4515
vīduiiūm: 4240
vīδuīūm: 4400
vīδoiiūm: 4410
V5.24b vīdōiiūm: 4200, 4210, 4230,
4420, 4510, 4515
vīduiiūm: 4240
vīδuīūm: 4400
vīδōiiūm: 4410
V5.25 vīdōiiūm: 4200, 4210, 4230,
4420, 4510, 4515
vīduiiūm: 4240, 4400
vīδōiiūm: 4410
Book 1.indb 148 24/05/16 13:44

Table 5c Combined Yasna

hōiiumca -um -ūm other endings
Y11.4 hōiiumca: 400 hoiiūmca: 415
haoiiumca: 420
Y11.5 hōiiumca: 400 hoiiūmca: 415
haoiiumca: 400
Y2.13 vīdōiiūm: 400
vīdiiūm: 415,
420
Y6.12 vīdōiiūm: 400 vīdūīm: 415
Y17.10 vīdoiiom: 400
Y25.6 vīdōiiūm: 420 vīdoiiom: 400
Y71.5 vīdōiium: 400 vīdōiiūm: 420

hōiiumca -um -ūm other endings
Y11.4 hoiiumca: 613 hōiiūmca 500
hūiiūmca: 510, 530
Y11.5 hōiiumca: 613 hōiiūmca 500, 510, 530
Y2.13 vīdōiiūm: 500, 510, 530
Y6.12 vīdūiiūm: 500
vīduiiūm: 510, 530

hoiiumca -um -ūm other endings
Y11.4 hoiiumca: 682
Y11.5 hōiiūmca: 682
Y2.13 [vīdōii]um: 677 vīdōiiūm: 681
vīδōiiūm: 682
Y25.6 vīδoiiom: 682
Book 1.indb 149 24/05/16 13:44

150 a. cantera
Table 5f Wīdēwdād Pahlavi

V1.8 harōiiūm: 4601, 4670, 4711, harōiiōm: 4700
4712, 4713, 4715 harōim: 4710
V9.17a hōim: 4600, 4610,
4670, 4700, 4711,
4713, 4715, 4710
V9.17b hōim: 4600, 4610,
4670, 4700, 4711,
4713, 4715, 4710
V9.18a hōim: 4600, 4610,
4670, 4700, 4711,
4713, 4715, 4710
V9.18b hōim: 4600, 4610,
4670, 4700, 4711,
4713, 4715, 4710
V8.25 ōim: 4601, 4700
aoim: 4610, 4711,
4713, 4715
V16.12 ōim: 4600, 4610,
4610, 4660, 4670,
4700, 4710, 4711,
4713, 4715
V20.4 aoim: 4600, 4610,
4670, 4700, 4710,
4712
aom: 4660
V5.22 vīdōiiūm: 4601, 4700, 47100,
4711, 4713, 4715, 4757
vīδuiiūm: 4670
vīduiiūm: 4712
V5.23 vīdōiiūm: 4601, 4700, 4711,
4713, 4715, 4710
vīδōiiūm: 4712
vīdaēiiūm: 4757
V5.24a vīdōiiūm: 4711 vīduuiiaīm: 4670
V5.24b vīduuiim: 4670
V5.25 vīdōiiūm: 4700, 4711 vīδuuiiīm: 4670
Book 1.indb 150 24/05/16 13:44

Table 6: daēum

-um -ūm other endings
V10.9 daēum: 4000, 4010, 4050, 4055
V10.13 daēum: 4010 daēuum: 4000, 4050, 4055
V13.6 daēum: 4000, 4010, 4050, 4055

V13.7 daēuum: 4000, 4010, 4050,
4055
V10.9 daeum; 4200, 4210, 4240, 4420, daem: 4230, 4410
4400, 4510, 4515
V10.13 daeum: 4200, 4210, 4240, 4420, daem: 4230
4410, 4400, 4510, 4515
V13.6 daeum: 4200, 4210, 4230, 4240,
4420, 4410, 4400, 4510, 4515
V13.7 daeum: 4210, 4240, 4420, 4410, daem: 4200, 4230
4510, 4515
Table 6c Wīdēwdād Pahlavi
V10.9 daēum: 4600, 4610, 4670, 4700,
4711, 4715, 4710
V10.13 daēum: 4600, 4670, 4700, 4713, daēm: 4610, 4711
4715, 4710
V13.6 daēum: 4600, 4670, 4660, 4700,
4713, 4710
daeum: 4610, 4711, 4715
V13.7 daēum: 4600, 4660, 4700, 4713,
4710
daeum: 4610, 4670, 4711, 4715
Book 1.indb 151 24/05/16 13:44

152 a. cantera
Table 7: *ºhṷam
hauuaŋhum -um -ūm other endings
Y71.11a hauuaŋhum: hauuaŋhəm: 201042
2101, 4000,
4010, 4055
hauuaŋvhum:
2010
Y71.11b hauuaŋvhum:
2101
hauuaŋhum:
4000, 4010,
4050, 4055
paŋtaŋhum -um -ūm other endings
Y19.7 paŋtaŋhum: paŋtaŋhąm: 4025,
4000, 4010, 4040, 4050
4055 paṇgtaŋhąm: 20
paṇŋtaŋhum: pataŋhąm: 2010
4060, 4100
paṇgtaŋhum:
2101
V6.32 paŋtaŋhum: paṇgtaŋhąm: 4010,
4000 4025, 4040
paŋtaŋhąm: 4050, 4055
V16.2 paŋtaŋhum: paŋtaŋhąm: 4000, 4050
4010, 4055
Y71.11a hauuaŋhum: hauuaŋhō.hūm: 231 hauuaŋhauum: 100
4420 hauuaŋhūm: 4510 huuaŋuhuum: 235, 234
huuaŋum: huuaŋhūm: 4515 haūm.aŋhumē: 230
4240 hauuaŋhauuaṇmē: 120
auuaŋhum: hauuaŋhō.mē: 4400
4410 hauuaŋuhē: 4200
Y71.11b hauuaŋum: hauuaŋhūm: 235, 234, 4420, haūm.aŋhu.mē: 230
4240 4510 hauuaŋhauum: 100
auuaŋhum: huuaŋhūm: 4515 hauuaŋhauuaṇmə: 120
4410 hauuṇŋhō.hūma: 231
hauuaŋhumē: 4200
hauuaŋhō.mə: 4400
42
Corrected into hauuaŋhum.
Book 1.indb 152 24/05/16 13:44


Y19.7 paṇgtaŋhum: paṇgataṇghūm: 4200, 4210, paṇɤa.tṇɤəm: 231
100 4240 paṇgtaŋhəm: 120, 234
pəṇgataṇghūm: 4250 paṇg.taŋhəm: 235
paṇgataṇgūm: 4420 paṇgataŋhəm: 4510,
paṇgataŋghūm: 4410 4515
paŋhe.taŋhūm: 4400
V6.32 paṇgtaŋhūm: 4200, 4210, paṇgtaŋhəm: 4230
4420 paṇatatŋhəm: 4410
paŋataŋhūm: 4340 paṇg.taŋhəm: 4510,
paŋhδaŋhūm: 4400 4515
V16.2 paŋhu.taŋhūm: 4400 paŋhutaŋhəm: 4200,
4210, 4240, 4420
paŋhutaŋhm: 4230
paṇŋhataŋ́həm: 4410
paṇg.taŋhəm: 4510
paṇg. taṇgəm: 4515
Table 7c Combined
Yasna
Y19.7 paŋhataŋhəm: 420
Y71.11a hauuaŋhəm: 400, 415
hauuaŋhē: 420
Y71.11b hauuaŋhəm: 400, 415,
420
Y71.11a hauuaŋhəm: 500
Y71.11b hauuaŋhəm: 500, 530,
682
Y19.7 paŋtaŋhəm: 500, 530
Y19.7 pgŋtaŋhum:
677
Book 1.indb 153 24/05/16 13:44

154 a. cantera
Pahlavi
V6.32 paŋtaŋhąm: 4711, 4715
paŋhtaŋhąm: 4601,
4712
paŋhaitaŋhąm: 4700
paŋhtaŋhəm: 4670,
4713
paŋhaotaŋhəm: 4710
V16.2 paŋtaŋhąm: 4600,
4660, 4711, 4713
paŋhąm: 4710
paŋtaŋhəm: 4700
paŋhataŋhəm: 4670
Book 1.indb 154 24/05/16 13:44

Bibliography
Cantera, A. (2012a). Building Trees: Genealogical Relations Between the Man-

uscripts of Wīdēwdād. The transmission of the Avesta. A. Cantera
(ed.). Wiesbaden, Harrassowitz: 279-346.
Cantera, A. (2012b). Why do we Really Need a New Edition of the Zoroastrian
Long Liturgy? The transmission of the Avesta. A. Cantera (ed.).
Wiesbaden, Harrassowitz: 439-478.
Cantera, A. (2013). Los manuscritos de Yasna con traducción pahlaví e instruc-
ciones rituales (abāg zand ud nērang). Séptimo centenario de los
estudios orientales en Salamanca. A. Agud, A. Cantera et al.
Salamanca, EUSAL: 503-521.
Cantera A. (2014). Vers une édition de la liturgie longue zoroastrienne. Pensées
et travaux préliminaires. Leuven, Peeters.
Cantera, A. (in press). “On Avestan text criticism (1): The importance of the
Iranian manuscripts for an edition of the Zoroastrian long lit-
urgy.” Iranische und indogermanische Sprachwissenschaft und
Onomastik. Akten des Symposiums in memoriam Manfred Mayr-
hofer (1926-2011), organisiert vom Institut für Iranistik und der
Wiener Sprachgesellschaft (Wien, 10.-12. Mai 2012). Schmitt,
Rüdiger - Sadovski, Velizar - Luschützky, Hans Christian (eds.)
Wien: Verlag der ÖAW. (Sitzungsberichte der ÖAW / Iranistische
Onomastik, Nr. 14).
De Vaan, M. (2003). The Avestan vowels. Amsterdam-New York, Rodopi.
Geldner, K. F. (1886). Avesta. The sacred books of the Parsis. Stuttgart, Kohl-
hammer.
Hoffmann, K. (1979). Das Avesta in der Persis. Prolegomena to the sources on
the history of Pre-Islamic Central Asia. Budapest: 89-93.
Hoffmann, K. (1975-1992). Aufsätze zur Indoiranistik. Wiesbaden, Reichert.
Hoffmann, K. and B. Forssman (1996). Avestische Laut- und Flexionslehre.
Innsbruck, Institut für Sprachwissenschaft.
Hoffmann, K. and J. Narten (1989). Der Sasanidische Archetypus. Unter
suchungen zu Schreibung und Lautgestalt des Avestischen. Wies-
baden, L. Reichert.
Kellens, J. (1996). «Commentaire sur les premiers chapitres du Yasna.» Jour-
nal Asiatique 284: 697-705.
Kellens, J. (1998). «Considérations sur l’histoire de l’Avesta.» Journal Asia-
tique 286: 451-519.
Kellens, J. (2006). Études avestiques et mazdéennes vol. 1. Le Ratauuō vīspe
mazišta (Yasna 1.1 à 7.23 avec Visprad 1 et 2). Paris, Éditions de
Boccard.
Kellens, J. (2012). Contre l’idée platonicienne d’Avesta ou les Considérations
revisitées. The transmission of the Avesta. A. Cantera (ed.). Wies-
baden, Harrassowitz: 49-58.
Martínez Porro, J. (2012) Análisis de dos fenómenos ortográficos vocálicos en
los manuscritos avésticos. Trabajo de Fin de Máster. Universidad
de Salamanca (unpublished).
Book 1.indb 155 24/05/16 13:44

156 a. cantera
Tremblay, X. (2012). Ibant obscuri uaria sub nocte : Les textes avestiques et
leurs recensions des Sassanides au XIIIe s. AD en particulier
d’après l’alphabet avestique. Notes de lecture avestiques VIII.
The transmission of the Avesta. A. Cantera (ed.). Wiesbaden,
Harrassowitz: 98-135.
Book 1.indb 156 24/05/16 13:44

AI57 Cantera

Uploaded by

Copyright:

Available Formats

AI57 Cantera

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AI57 Cantera

Uploaded by

Copyright:

Available Formats

This pdf is a digital offprint of your contribution in E.

Pirart (ed.), Etudes de linguistique iranienne in memoriam

The copyright on this publication belongs to Peeters

As author you are licensed to make printed copies of the

If you wish to publish your article immediately on open-

For queries about offprints, copyright and republication

études de linguistique iranienne

Contributions rassemblées par

98106_Acta Iranica_57_vw.indd 3 24/05/16 13:39

98106_Acta Iranica_57_vw.indd 5 24/05/16 13:39

Book 1.indb 89 24/05/16 13:44

Keywords: Avestan grammar, Avestan transmission, Avestan manuscripts

In Hoffmann-Forssman’s (1996) description of the forms of accusa-

Book 1.indb 90 24/05/16 13:44

Book 1.indb 91 24/05/16 13:44

The manuscripts checked are the following:

Book 1.indb 92 24/05/16 13:44

The accusative singular of the u- stems

Book 1.indb 93 24/05/16 13:44

The reading ahūmca ratūmca is a lectio facilior common only in India.

Book 1.indb 94 24/05/16 13:44

Nevertheless, in Y11.3 not a single variant with ‑um is found. Together

Book 1.indb 95 24/05/16 13:44

1. -ū instead ‑ūm appears almost exclusively in the Indian manu-

Book 1.indb 96 24/05/16 13:44

The accusative singular of the ū- stems

Book 1.indb 97 24/05/16 13:44

as to fsəratū-. The ending appears as ‑ūm in all manuscript classes,

The accusative singular of the ṷa- stems

Book 1.indb 98 24/05/16 13:44

The Iranian manuscripts perfectly distinguish between the acc.sg. of

Book 1.indb 99 24/05/16 13:44

Thus there is no doubt that the Iranian manuscripts still distinguish

– pourum (acc.sg. m. of pouruua-) appears thus in all passages V4.47,

Book 1.indb 100 24/05/16 13:44

Dārāb Hīrā Candā as 4210, shows a tendency towards paourəm

Book 1.indb 101 24/05/16 13:44

occasionally as early as in Mihrābān’s manuscripts, especially in the

Book 1.indb 102 24/05/16 13:44

Iranian Indian Mihrābān combined other Yasna

Book 1.indb 103 24/05/16 13:44

Book 1.indb 104 24/05/16 13:44

paŋtaŋvhūm -um -ūm (or ‑uum) -əm/-əm -ąm

Y71.11a hauuaŋhum: hauuaŋhō.hūm: 231 hauuaŋhəm: 400, 415

Book 1.indb 105 24/05/16 13:44

This frequency of -um/-ūm never occurs in other final words with

The development of the acc.sg. of these stems

Book 1.indb 106 24/05/16 13:44

Iranian Indian Mihrābān combined other Yasna

This distribution can be explained as the result of a series of evolu-

Book 1.indb 107 24/05/16 13:44

group of manuscripts disyllabic words ending in ‑ōiium were reduced

Book 1.indb 108 24/05/16 13:44

Book 1.indb 109 24/05/16 13:44

Book 1.indb 110 24/05/16 13:44

gaōdaiiūm -um -ūm other endings

Book 1.indb 111 24/05/16 13:44

pitūm -um -ūm other endings

Book 1.indb 112 24/05/16 13:44

– pourum (acc.sg. m. of pouruua-) appears thus in all passages V4.47,