The inherent vowel in the alphabet of the avesta-language.
By L. H. Mills.
In view of certain propositions which were made by me al¬
ready in the Sacred Books of the East, Vol. XXXI (1887), and
continued in the commentary to the Gathas pp. 394—622 (1892
—94), also in a few short articles in this periodical, I trust that
I may hope that disinterested scholars will continue to agree with
me in my general suggestions as to the zend alphabet, details
being often of course only tentatively advanced.
It is quite obvious then that certain of our zend characters have
been misunderstood, and in the hope of engaging the interest of those
who are specialists in zend phonology, I will dwell for a moment on
some obvious peculiarities of the avesta alphabet before discussing fully the inherent vowel. The sign f{j for instance, is not at all properly
explained as being in any sense whatsoever, the equivalent of the
sound f in some of its uses. In its application, for instance, to
express the gen. s. m. of the it (a) declension it is clearly an ancient
sign retaining its original force, which was totally distinct from
that which a similar character most often possesses in the zend
alphabet.
In the use to which I refer, that is to say, to express the
last syllable in the genitive form of the Ji or a declension, i'' Lias
nothing whatever to do with tbe sound ?, and is altogether errone¬
ously replaced by such a transliteration (Darmesteter among others
following me in this view).
The sound g, or anything like it, is totally absent from this
= sk. -asya.
No such genitive singular masculine of the a, a, declension
as -ak? ever existed, or was in any way thought of, by those who
Bd. LV. 88
344 Mills, The inherent vowel in the alphabet of the avesta-language.
spoke the language of the Avesta as their vernacular. The character
^ (j^) does not merely represent y -\- a, (-ya) here as in -ahya
(sk. -asya), gäthic -ahyä, but it is here the two characters for
y -\- a themselves as usually cursively united as in the well-known
pahlavi sign = ya in the original avesta-pahlavi writing.
^ (^) is simply 0(3 accidentally or purposely lengthened,
and (') is actually and literally avesta-pahlavi i» + j cursively
written .»(3, J being the original mark for y (among other sounds)
iri the pahlavi of our MSS., and in pahlavi equalling long ä
(also at times initial short a).
Another example which I cited as striking was ^|jjj^, in
which word the sign f{) is again obviously the ancient repre¬
senting a -\- i cursively written as it is in the pahlavi of our
extant MSS. ^ being the same J^j accidentally or purposely
lengthened perhaps especially to show more definitively a final
long a =: ä instead of a short i> (a). And this also especially
lengthened J^j again does not merely represent Ji -4- i = y ^ a,
but it actually is the graphic combination of those characters, the
word is Jcainyä, cf ind. kanyä (so), n. s. f. of the a declension. I
may add another more striking instance of this misuse in these
preliminary remarks (as in passing). It occurs in that sg.
imper. ^.i)jj| , which, as no one doubts, in some way represents
nasya. But this character ^ not only stands (in some way) for
j» -f i = ya (or yu), but it is (again) the actual original ancient
avesta-pahlavi writing of the two letters, avesta-pahlavi J = later
full avesta JO (= ?/)-)- J» = a or 3; it is old avesta-pahlavi
accidentally or purposely lengthened to Nas^ (sic) is a
so-called "monster". To introduce the letter p in connection with
^ as here used, is simply a negation of fact. No such sound
1) Those who do not read zend constantly are reminded that ^(j is
chiefly used for s in the fully developed avesta alphabet; but so also in the
previous pahlavi = y a and also s, etc.
as e was ever in any way used (or thought of) in this application by speakers of a vernacular.
But there is another usage which needs explanation quite as
much, though from a somewhat different point of view ; and it is
one which illustrates the presence of the "inherent vowel" in the language of the Avesta.
There are some characters which we are in the habit of trans¬
literating correctly from the zend into latin, but which are them¬
selves, as clear zend characters, erroneous transliterations of the
earlier and quasi original avesta-pahlavi alphabetical signs.
Take for instance, the sign ^ = ft in the termination of the
2°'^ pl. imperative med. of our gathic verbs. Surely no one will
doubt that these forms are wholly disfigured by being represented
as fi^^ , düm. d, — ü, — m, are indeed sufficiently ample reproduc¬
tions of the signs £ ,— (J ,-— J , but no one should any longer suppose
that fi^ , düm, really represents any sound that was ever spoken
by an ancient Iranian as a syllable expressing the termination of
the second pl. imperative, optative or preterit middle of any word.
The man of the gathic age never said:
i^zJi'^i^ i/C'-ez-t^wwJ (?)
g^^-^J^^- -vöi'z-düm (?)
>^*(jj ^Qg_ (/üshö-düm (?)
-^juj)c^ thräz-düm (?)
- säz-düm (?)
and the like, while even the men of the later Avesta said:
Ha^(^m»)ti>^ darayadlmem ')
^^^(ouiiJ)iul^ värayadhwem,
the indians in tbe mean time having never varied with -dhvam, as
in dharayadhvam, trddhvam, Jushddhvam, dadhidhvam, etc. And
it must have been long doubted that fi^^ , düm was genuine ; what
I am offering here is an explanation of the phenomenon.
1) w is for Englisli u; and in fact the Indian interior half-vowel should strictly speaking be often transliterated w not v; elsewhere otherwise.
23»
346 Mills, The inherent vowel in the alphabet of the avesta-language.
We decide at once that the reported form , düm, is a
mere error in writing and of comparatively late origin; and that it
has nothing to do with any original sounds of the language in this
place. But how did it arise?; and why was it perpetuated? I
think that it is very evident how it arose. The original consonant
of the avesta, like the Sanskrit and old persian consonants and
some of the half-vowels, and still more, like the more modern
pahlavi (the mother-schrift of the avesta) possessed an inherent
vowel. In the Sanskrit it is simply a, but in our pahlavi MSS.
all the short vowels are inherent in the consonants, save only that
initial short a is reproduced by the same sign JJ which expresses
long a, whether initial or otherwise.
The character which was mistaken by those who perpetuated
the gathic -düm was the original avesta-pahlavi sign which
developed into the form of ) in our pahlavi MSS. here imitated
in the only pahlavi types which we at present possess. (Let it
be remarked that I do not say that the sign was ), but that it
was that of which ) was the result. ) was the outcome of con¬
fusing different, though doubtless somewhat similar characters, as
has happened in the case of many others of these ancient marks ;
and I have personally no doubt whatsoever that in many cases the
original characters which became united in forms such as this were
intentionally so confused for the purpose of establishing a writing
only intelligible to the initiated.)
This quasi original avesta-pahlavi character ) represents several
different sounds in the pahlavi writing of our present MSS., and
among other sounds it would undoubtedly represent the english w,
as also the closely related spirant v. The use of this ) to express
the w sound in the 2°* pl. med. imperativ, optative and preterit,
was therefore not only appropriate, but very natural, though not
positively necessary. It was not the only sign for the v or for
the w sound, but it was the simplest and the shortest. When
then the ancient copyists wrote (say) or something
similar, they were completely in order so far as ) was concerned,
though the use of the more newly developed signs for the rest
of the word betrays a transition period, (quasi) original avesta-
pahlavi ) very properly represented the (english) w sound in the
2""^ pl. imper. med., but only when understood as jiossessing an
inherent full vowel sound, though being itself a semi-vowel. The
i ] of well expressed the CJoi'(= wem) of iliX^(^ =-dhwem
(though the ^ ^ in place of (dh) shows traces, analogous with
those in ), of being a more rudimental form for the fl = dh, the
rationale of which I will not discuss here.) ^
I am for the moment only concerned with ). This at one
stage legitimate and, so to speak, original avesta-pahlavi | , well
possibly = english w as well as v with its inherent full vowel
sound e or a, as english we or wa, having retained its place as
a relic of antiquity in the transcription of the 2""* pl. imper., pret.
and optative middle >) in the oldest part of the Avesta the Gäthas,
became in time misunderstood, as was only natural.
Why should a more original avesta-pahlavi sign | be retained
in the middle (?) of a word made up of fully developed avesta
characters, the clear and copious sehrift invented, say, so late as
the third, fourth, or fifth century after Christ (perhaps the plainest ancient alphabet of Asia). The simple copyists, dissatisfied and puzzled, began to transliterate (?) this sign, as their predecessors had long since
transliterated the signs which were the originals of
the other signs in the word (whatever those original signs may
actually have been).
And what more natural choice could they have made for
their blunder (if they were destined to make a mistake at all) than
to choose = Ö ?; for = ü was and is one of the most
natnral sounds among the several differing ones which in course
of time had become clumsily included as represented by the quasi
original avesta-pahlavi sign J ; and the result was our "monster"
^ = w in = düm which disfigures the otherwise so often')
superior writing of the Gäthas.
This ^ = is wholly out of place in the spelling of the word
so long as we possess a ^jii^we-) or a J» we-). No ü sound (pure
and simple) was ever uttered in this connection (as not having
any other accompanying vowel) by a native Iranian whose verna¬
cular was the language of the Avesta.
The letter ^ = « should be as scrupulously avoided or ex¬
plained here as the spurious p of our faulty transliterations, distor¬
ting and falsifying, as it does, the ancient grammar, as we have
seen, by reporting a termination -ah(} (for the gen. s. m. of the a
declension) which termination never carae from human lips while
1) not that other instances of defective transcription do not appear there.
2) as to tW'and », english v for Wi'nnd engl, w for » are convenient for usual application , but I do not think that this explanation holds without e-xception; see Whitney on ind.
348 Mills, The inherent vowel in the alphabet of the avesta-language.
uttering their mother speech, fj should be replaced by the signs
wbich alone could have any pretence to express the full sound in
this syllable of the language; that is to say, unless we wish to
persist in the obscuration of the facts. We should write oj-
*ȟ/(or or everywhere when representing the last
syllable in the real Iranian 2°'' pl. imperative, optative or preterit
middle ; the words are :
<citi^*^^i^l^ gerezdh*wem
i{a^*(S^i^l^ vöizdh*werr>.
€J«J/*(»^J2J|^ güshödh*wem
Hti^^^**>)<!} tliräzdh*wem
säzdh*wem, etc.
Cp. again ind. -dhvam (sic) in trddhvam etc., as above; but more
especially the new avesta GJO^ß- {-dhwem) in CjOi^Q^JJjijfiJ^
eembayadhwem, etc. (see above).
It is then not conceivable that the contemporary of the Gathas
said C'f^ ^){(^ gerezdüm and the like, while the composers of the
later avesta said gjtx^(Oijjjii)jju^ , därayadhwem, etc.
The character r= a used in this connection is a mistaken
transliteration for a more original | (or for its predecessor). And
this mistake in reproducing ) with a single letter ^ instead of
with (j)8X^with an expressed vowel shows that, as in the more
original avesta alphabet and as in our pahlavi, this ) itself, before
its false transliteration into ^ = «, expressed without any auxiliary
sign an inherent vowel necessary to make the sound represented
to be uttered intelligible to the hearer as the above mentioned
form of speech, viz. as the 2°'' pl. imperative, preterit and opta¬
tive med.
Another familiar instance of this ancient error occurs in tiie
reported form fi^^, turn, for the pronoun of the second person
.singular. There is no question at all but that the word intended
to be reproduced was fij»^ english trcem; cp. gäthic =
english twem, indian tvdm (sic) spoken english twam, and tudm.
There was never any such a word as tüm = "thou" in either
Sanskrit or zend. The word was of course originally (or better
'previously') written C ) ^ , the J being, as in the case of the spurious
verbal termination treated above, a previous avesta-pahlavi sign
expressing oi'or » = english w, (as also english v, with other
sounds) ; and like all the pahlavi consonants or half-vowels it con¬
tained its original full vowel sound, in this case J = e (or i>
= <^)) fi ) ^ representing tw{e)m (or tw{a)m). But, as in the cases
mentioned above, some simple later copyists found it necessary to
abolish the old sign J as their predecessors had abolished most of
the other quasi original avesta-pahlavi signs, replacing them by
the characters of the fully developed avesta alphabet. And in
choosing a character which was not the right one, they again fell
upon the very natural ^ = ü, giving us our impossible ,
tUm,^) with no expressed e, or a vowel, it being left as inherent
in a mistaken fellow vowel as tu{a)m, or else completely lost to
sight through lack of scholarly capacity.
To illustrate the matter a little further, I cite a few analogous occurrences.
G^öy "J**"* (') should be restored as gji»j»y^2) (or GJ»-),
javam (or -vem), cp. ind. javdm.
bün* (?) should be restored |i>>>i)^ bavan, cp. bavän.
k^istüm (?) should be gii»^Ji(j{^ (or C{»-),
hhstvam (or -vem), cp. khStvo.
fi^^ drum (?) should be fiJl»))^*) or fiJ» -) druvam (or
-vem), cp. dr{u)vö, ind. dhruvam.
1) as regards the exceedingly interesting and important gäthic and later avesta (tiT), >^ tu, it seems hazardous to suggest a doubt of them; but, the question must nevertheless be asked whether an ancient sign of nasalisation has not been lost from them. Are they not in reality '^V, '><( for cp.
ind. signs of the nasalisation having been often lost in the Avesta.
2) perhaps more originally (?) = -vem everywhere instead of = -vam.
3) = a is twice inherent; ) (or its predecessor) here represented the spirant v.
4) the first is not (?) so much a latent inherent vowel as simply an omitted sign owing to the following ».
350 Müh, The inherent vowel in the alphabet of the avesta-language.
C^j^j)(^ thrüüm (?) should be ia)>f^i)^ (or g^»^
thriSvam (or -vem), cp. thriSva.
The forms in their defective state are wholly misleading; they
are each and all mistaken attemps to reproduce a previous J
(or its original), and this mistaken reproduction omits to express
the originally inherent vowel.
A person to whom the language was vernacular would not
have understood what was meant by the sound, if "ÄöJi" (or "■bun") were said to him instead of ''bavan".
Sometimes the transcribers, possibly influenced by a foregoing
Jl (= a), chose the letter i as the interpretation of the quasi
original avesta-pahlavi ) (or for that sign which it superseded),
for, as they did not provide this X with the pendant stroke as
the sign of prolongation, 1> was even a more natural reproduction
of ) (or of the sign into which ) developed in our MSS) ; and in
so doing they again negflected to write the inherent vowel jj (= a) ;
cp. for instance :
jijjigXji aomana (?) which should be j>j*JUgjJ»ji ') ava-
mäna, cp. J^JJJ«»ji, aväiti, etc.
j»JgXji)fi, -mraomna (?) which should be Ji)*J>i6J>»Jt)(; '),
•mravamäna, cp. .^^JJ>»Jt)c, mravaiti, etc.
jXjJ^ baon (?) which should be |j) »ju j , bavan, cp. dbhavan, etc.
jijgXjj^ daomna (?) which should be ji|*JuCJi»jt^ i), dava-
mäna, cp. ji^J»»JJ5Ji, adavata, etc.
gXjt^iO yaom (?) which should be (or C{»-) ^^(»Ji)«^, yavam
(or -vem), cp. ■^»^)*(5, yavö, cp. ydvam, etc.
gXjJJjjJJtC mainyaom (?) which should be CJi»Jtjj|jji£
(or £(»-), mai'nyavam [or -vem), cp. mainyava, etc.
1) an omitted or shortened long ", ä, should be discussed elsewhere.
Finally, and as perhaps the most natural usage of all, in¬
fluenced by considerations which we need not state definitively here,
they sometimes chose the sign > used in the later alphabet for u
to express the ) (or its predecessor) as equalUng v with an inherent
vowel ;
paourum (?) should be (or fii»»)>l)jj , (or fiJ» -),
jjaourvam, (or -vem), cp. paourvo, ind. pürva.
|>j«j baun (?) is to be restored as ^) |jj>>j>| bavan, cp. {d)bha- van, etc.
bun (?) to be restored as Jj»>>iiJ , bavan, cp. dbhavan, etc.
Jt^^>JU^i> adaunta (?) to be restored as*) Ji^^J»>Ji^i>
adavahta, cp. adavata.
6>)>JHy haurum (?) to be restored as CJt»))ji^ (or £{»-))
haurvam (or -vem), cp. sdrvam.
C>iU^^-, -fshäum (?) to be restored as ^^»ü^^e)- (or -))
-fshavam (or -wm), cp. -fshav^.^)
1) The short j,, a, is also to be restored as the sign; for the lengthened sound J. ä was evidently intended to express accent, but was doubtless at times incorrectly used.
2) There are very many other instances of the inherent vowel in the avesta language, many irrational variants having resulted from unsuccessful attempts to reproduce tbem. See for other instances this Zeitschrift Bd. 49, III. Heft, 581, 1895.
352
Berichtigung zu S. 136.
In dem interessanten und auch für die Sprache des Talmuds
lehrreiche Bemerkungen darbietenden Aufsatze von B. Jacob über
.Christlich-Palästinisches« (ZDMG. Bd. 55, S. 136) findet sich ein
bedauerliches Versehen in der Wiedergabe einer TalmudsteUe, das
hiermit berichtigt Vierden möge. Jacob sagt: „Er — der babylo¬
nische Talmud in Pesachim 7 b — fragt iNni »ttiaw und
antwortet Nirt ^unpNT t<:\a"'b''. Die citierten Worte aber, die von
Jacob als Prage und Antwort betrachtet werden, bilden einen
einzigen Satz und stellen bloss die Prage dar: „Was macht es er¬
sichtlich , dass jenes laiy — in dem vorher citierten Ausspruche
des Amoräers Samuel — ein Ausdruck für das Vorhergehende ist?«
Die Antwort auf diese Frage wird dann von späteren Amoräern
mit Heranziehung der von Jacob citierten Bibelstellen erteilt.
W. Bacher.
Berichtigung zu S. 145 ff.
Ich bedaure sehr übersehen zu haben, dass bereits L. Stern
im 23. Bd. der Ägypt. Zeitschr. S. 119 die ägypt.-arabischen Frage¬
sätze richtig erklärt hat. Herr 0. v. Lemm hat mich darauf auf¬
merksam gemacht. F. Praetorius.