J
\
GIFT OF
Benjamin Ide Wheeler
<2A* '//?*
THE
AVESTAN ALPHABET
AND ITS
TRANSCRIPTION
BY
A. V. WILLIAMS JACKSON
OF COLUMBIA COLLEGE, NEW YORK CITY
WITH APPENDICE
STUTTGART
W. KOHCHAMMEB
1800
THE
AVESTAN ALPHABET
AND ITS
TRANSCRIPTION
BY
A. V. WILLIAMS JACKSON
OF COLUMBIA COLLEGE, NEW YORK CITY
WITH APPENDICES
STUTTGART
W. KOHLHAMMER
1890
K(r l&l
-J 3
rAfi i a}
CONTENTS.
Avestan Alphabet and Transcription.
Page
The Avestan Alphabet : Its Character 5
Lack of Uniformity in Avestan Transliteration 6
A proposed Transcription of Avestan 9
Discussion of the Avestan ^-vowels 10
A Modification in Transliteration suggested 1 1
The Av. Consonants: Germanic Letters as Symbols 13
Diacritical Marks on Letters: Discussion . . 16
Summary of the suggested Transliteration 22
Appendices.
Proposed Transcription 26
Suggestions to the Printer 27
Substitute Alphabet 28
Systems of Transcription 29
Bibliography . . . , % 34
411144
*
The Avestan Alphabet
and its
Transcription.
What the original alphabet was in which the Avesta
was written we do not know. The alphabet in which our
texts are now preserved bears the stamp of a much later
age than the language it presents. The question of the
origin of this alphabet in which our MSS. are written has
difficulties ; but there is little doubt that it is derived from
the Pahlavi alphabet of the Sassanian times ; it is closely-
related to the book-Pahlavi. This point must be con-
stantly borne in mind in discussing the letters. The ques-
tion, moreover, of the transcription of this Avestan alpha-
bet as we now possess it has long been and is still a
very vexed one. This is the question, in particular, that
forms the chief subject of inquiry in the present paper;
but all investigations into the matter of transcription imply
a more or less extended study of the alphabet from the
standpoint of palaeography, phonetics and philology.
The lack of uniformity in the system of transliterating
the Avestan characters is confusing to those pursuing philo-
logical studies; it has doubtless also in some degree re-
tarded the advancement of the Avestan cause. From the
standpoint of philology, the present necessity of some
accordance in the method of transcribing this language is
apparent. The time is not far distant, it seems, when
scholars should and will tend toward adopting some uni-
form system. Geldner's new edition of the Avestan texts
has practically fixed the number of characters to be tran-
scribed, and should Iranian students now agree — and it is
hoped that those in America may perhaps set the example
— in adopting some uniform method of transliteration, that
shall be practical as well as scientific, an additional impetus
would be given to these studies. Such adoption would be
a grateful service to all, particularly to those interested in
the linguistic importance of the Avesta to Philology. To
write on the subject may not be a thankless task ; if some
suggestion or hint thrown out lead but a step in the right
direction as a guide to others for finding a better way,
the labor will be quite repaid. To the linguist, moreover,
the comparative table of the various systems of transcrip-
tion, appended for reference (see Appendix), may not be
unacceptable.
In regard to the method of Avestan transliterations,
the number of systems is almost legion. Many of them,
however, differ from one another only in some minor
points; in fact, on most of the ordinary details there is a
growing tendency more and more toward uniformity. It
is chiefly in a certain few respects — but these points are
important ones — that Avestan scholars still mainly disagree.
Some of these differences have been due to variations in
the Avestan characters of some of the MSS. or to diffe-
rent forms adopted in the editions; but since the new
edition has set up a standard, the question of the actual
Avestan characters to be transcribed has become practi-
cally settled , and it seems as if greater agreement might
be brought about. Of course those who have to deal with
palaeographic questions of the MSS. will be compelled to
add other signs in transliterating, but this need not con-
cern philologists generally. By a few mutual concessions,
uniformity and concord in rendering the symbols of the
Avestan texts might soon result.
The transcription here offered is presented in a ten-
tative way, in the hope that some of the hints may prove
useful for the future. It has been based on personal ad-
vice and suggestions upon various points, from names of
no less authority -linguistic , palaeographic , philological,
and phonetic — than Professors Brugmann, Geldner, Pischel,
and Sievers. To these was added weight from the stand-
point of epigraphy — Professor Andreas. Practical sugges-
tions have also been received from Professors Delbriick,
Collitz, Hopkins, and Lanman. The marshalling of such
names is of itself not without significance; the question
is one that really is of interest to many scholars. The
opinions on the subject of course varied. The translitera-
tion, which I here suggest, is given as a sort of com-
promise and concession both to the radical and to the
conservative side of the question. The system has en-
deavored to be at the same time strictly scientific and yet
as far as possible practical. With a little good will, per-
haps out of this system some uniformity of method might
be developed and adopted. In America at least we have
now the opportunity of uniting; if a few will take the
lead, others will follow.
In preparation of this system the various methods of
transliteration (Bartholomae, Hubschmann, Justi, de Harlez,
Sacred Books, etc.) have been examined : the aim through-
out has been to hold the mediant viam. — The main fea-
tures of the system are (i) that it shall be scientific and
at the same time fairly practical. (2) Single characters
as far as possible are represented by single signs. This
latter is far more practicable, and at the same time more
requisite, in Avestan than it is in Sanskrit. (3) It makes
concessions as far as possible to existing systems, and
as far as may be avoids radical alterations and intro-
ductions. — The particular points characterising the
system are: (1) a remodelling in transcription of the
8
<?- vowels; — (2) adoption of the more or less generally
used Germanic characters d, ]?, j, // for the spirants;
similarly also for the nasal, cf. also aspiration; — (3) uni-
formity in the use of a diacritical sign to differentiate
letters. For such a diacritical sign in Avestan, the prac-
tical proposal is made to use a subscript tag % t (inverted
apostrophe, comma, spiritus lenis or the like) to designate
the Avestan 'derivation stroke' discussed below p. 16.
In regard to the character of the transliteration as
being scientific, it must of course be said that when ex-
tensive Avestan printing is to be done some of the tran-
scription types would have to be cast. But in most offices
that do philological work, the majority of the types re-
quired are already on hand ; the few that may not be, can
always be mechanically made without much difficulty. In
this way the practical side of the question has been kept
in view. In smaller articles the transcription can always
be used without the necessity of having the type cast;
the signs in general are such as can be made up by any
intelligent compositor. 1 In this respect the more consistent
use of the modifying 'tag' ( t t ) is very practical. The tag,
moreover, when cast on the letter breaks off less easily
than the point. For purely popular articles Justi's trans-
literation somewhat remodelled may of course be retained
— see Appendix 'Substitute Alphabet'.
The Av. transcription tentatively proposed is now
given. The order of letters is based on that of the San-
skrit alphabet. For convenience of reference, moreover,
Justi's transliteration, Handbuch der Zendsprache, is given
beneath each letter. The letters in parenthesis ( ) show
where deviations from Justi have been made.
1 Reference, for example, might here be made to several Notes by
the writer in the American Journal of Philology 1889 — 90 where the
system was thus employed.
9
Proposed Transcription of Avestan. 1
(Compared with Justi, Handbuch der Zendsprache).
A. Vowels.
Short *> a * i > u * d »o e ^> o
a i u (e) (e) o
Long -»" a y 2 ^^ 5^ ^ ^ > J p«<^ ^^
« ? £ (i) i d (do) (a)
B. Consonants.
Guttural ^ k fa fy ? g ^ j
k (kh) g (gh)
Palatal y C — ^J _
c J
Dental y> t 6 }> _j d ^d £ /
t (thj d (dh) (t)
Labial <u p & f _j b on It'
p f b w
Nasal } P as V I n ^fl -dm
(n) (n) n (n) m
Semivowel and
Liquid rc> («) y (t) * ^ V Is (») V (u) 2
y r v
Sibilant <&S -v s \& $ s rv $' d f z <&) Z
(() (0 (**) (sk) ■ z (zh)
Aspiration . . . . <y h \» J{
h (a)
Ligature H" h r
(q)
In the above alphabet a certain number of the Avestan
characters are simple and have in general corresponding
Latin letters that may represent them. All unite now in
transcribing these in the ordinary manner. They are —
1 For a merely practical transliteration to be used for popular pur-
poses, see 'Substitute Alphabet' p. 28 below.
2 The signs i, u need be employed only for purely scientific pur-
poses; the letters y, v for both initial and internal ro ■", I? », answer
fully for practical purposes.
3 The single sign I is quite sufficient for the three T), £), ro.
The differentiation s, /, / need only be made in scientific articles.
10
a, a, i, i, u, u, o, o
k, g, t, d, p, f, b
n, m
r
s, z, h.
In regard to the others , questions of greater or less im-
portance arise and there is more or less diversity in re-
spect to them. Beginning with the vowels these may now
be taken up in detail.
A. VOWELS.
A Modification in Transcription of the Vowels.
Av. {, j — to, (U — ***»; >C.
<?, 5 — e, e — 00; q.
^-Vowels.
i. Av. { 3 (Justi e).
Av. .3,o^.u,ta kBranaoiti (Justi kerenaoitt).
The fact that Av. { is not a pure ^-vowel has long been
recognized. There have consequently been various different
methods of denoting it, e. g. ce or a (Rask, Ueber d. Alter u.
Echtheit der Zend-Sprache), ^(Spiegel, Av.-Uebersetz. etc.;
Geiger, Handbnch der Aivestasprache), e (Fr. Miiller, Wiener
Zeitschrift), etc. A critical study of this letter shows that
it must have represented an obscure sound that seems
closely to have resembled the short indefinite vowel familiar
in English, 'gard£n^r', 'history', 'potato', 'measuring', 'mut-
ton', '# gainst', 'forward', 'sachem', 'formidable'. It seems
to have approached sometimes more nearly to a (i), some-
times to ?c (5). 1 A study of the MS. variants in this re-
spect is very instructive.
The ^-tendency of $ is seen, for example, in endings
e. g. Av. -faiy^w yasmm, ^ baran = Skt. yajhdm, dbharan.
1 Cf. e. g. Ys. 57.2 variants pivorsstd, pwardsto, pw?rdsato, pwardsatd,
etc. Cf. also \ a below, GAv. n$ = YAv. no etc.
1 1
So GAv. avisti = Skt. dvittl} Again, observe the inter-
change between d and a in the MSS., e. g. upznwm beside
up-a-mzm, mainyavitn beside mainy-a-vo etc. Furthermore,
the 3 that arises from a before m or n may be palatalized
to i when either y, c, j, or z immediately precedes. The
variants show the palatal nature of the sound e. g. Av.
^*pj-»^ vacim beside -fap-"^ vacant 'voice' = Skt. vacant —
cf. Jackson, Avesta Gram. §§ 29, 30. Independent variants
between d and i occur quite often, even when no palatal
precedes, e. g. nantata-, nimata- Vd. 5.38, bizatigra-, ba-
zangra- Ys. 9.18 and many others.
The ?/-nature beside its /-shading is also notice-
able. In GAv., for example, d appears sometimes to be
written (as a kind of dissimilation) for u or i, when in the
following syllable an u (v) or an i stands; the epenthetic
vowel may then be also written beside it. This proves
the labial character in addition to its palatal. Thus, GAv.
-k^-"))^ drdgvaiit- 'wicked' (— *drugvatit- to Av. druj-)\
GAv. bdzvaiit- 'advantageous' (= *buzvant- to Skt. bhuj-)\
GAv. n$9 u ru- 'zeal' (?) Ys. 34.7, cf. u§-u-ruye Ys. 32.16 ; GAv.
hn&Hi- 'well-being' (i.e. hu$-i-ti-)\ GAv. dnaHi- Ys. 30.11;
GAv. askzHi- Ys. 44.17. See Av. Gram. § 31.
Such interchanges with a, t, u (0) are indicative of
the intermediate character of the sound. 2 For all these
reasons, it is here suggested to adopt the transcription (9)
commonly used in Phonetics (cf. Sweet's Hist, of Eng.
Sounds 2 ed. p. 15 etc.) for the sound that seems so closely
to resemble it. This choice of the 'turned e (a) is a
practical one; no new type need be cast. The compositor
has simply to reverse the e (a).
1 Cf. Eng. avowal, pron. pvowpL
2 Palaeographically Av. f is connected with Pahlavi 1 u, i. e. merely
a broken I ({). For some excellent points bearing on the Pahlavi and the
Inscriptions in regard to this letter, I am personally indebted to the cour-
tesy of Professor F. C. Andreas.
12
2. Av. j £ (Justi e).
GAv. ^/" a-s^z, {i m, Av. *&>},? g$us (Justi 0£/*0, »^ £^).
The sound $ * is the corresponding long to j * and is
therefore to be represented (9~) in accordance. It is found
chiefly in GAv. 1 answering to YAv. 1 9, a, 3, q. 2
3. Av. fo e, ^ e (Justi e).
GAv. ^>*>$-»w>yazaite, YAv. y^^y^yazaite (Justi yazaite).
These two, to «, correspond to each other in the MSS.,
as short and long. They are therefore to be distinguished.
Justi etc. in accordance with the first editions gave both
as e. Later it became customary, as was proper, to dis-
tinguish them from one another and a subscript point
(thus e e) was used to differentiate them from e e which
were adopted for j j. This now is no longer necessary;
as we have 9 9 for $ j, the simple e e for to (o may be
adopted. That brings them in direct accord with their
parallels"^ 0, > o. z
4. Av. r*> m (Justi do).
Av. ^^3/-"^ mazdm (Justi mazddo).
Palaeogmpjikally, p« is evidently a combination of
t~>d9. Phonologically, it seems to have denoted an in-
1 GAv. = Gatha Avesta, all that is written in the Gatha dialect—
YAv. = Younger Avesta.
2 Phonetically \ 3 probably resembled the long drawn English pro-
nunciation of 'w^rd' (waard i. e. or), Varth' (aarth i. e. er), 't«rf (taarf i. e.
ur), 'f/'rst' (faarst i. e. ir) etc., cf. Sweet, Hist, of Eng. Sounds p. 276.
3 If "^ "r 3 are universally rendered o, consistency requires that to (0
should likewise be given without the subscript point. Perhaps both sets,
however, would orthographically better be given by some diacritical mark
e. g. 0, 5, e, e (though (.) generally denotes a lingual letter). They are
not pure sounds. This is shown, for example, by vohu, cor a £, yesnc, and
such MSS. interchanges e, a, i etc. It must be remembered that Av. o, e
= Skt. d, e only when final. Ordinarily Skt. 0, e are represented in Av.
by ao, ae. A fuller discussion must be reserved for some future time.
13
definite ^-shading of *» a. It fluctuates in the MSS. on the
one hand with ■*« a e. g. -«^/-»^ mazda as variant to ^/-"-G
mazdm; on the other hand it is found as a variant for
)-»« au e. g. ^ro-"V»- lp'atc& for >^ro^^ fyratdu. The palaeo-
graphic and phonetic character of the letter, then, are
thus given
^ ( iM ) = & (da).
In printing, <^ if not provided, may be mechanically made
by uniting a a under the macron cm.
5. Av. )c a (Justi a).
Av. -foc<y /z#7/z, wc»^ daevqn (Justi /jtfV/z, daevdri).
The letter ^ presents a nasalization of a, a; sometimes
it also resembles Skt. anusvara. To render it, the cha-
racter q, has been preferred to a. The symbol q is now
more generally used in linguistic works to represent the
tf-sound with nasal coloring. It is preferable in case the
question of accenting the vowel (') ever comes up. Of
course in popular works a may be retained. The symbol
q may be me chanicall y made by placing a subscript hook c
below the letter. 1
B. CONSONANTS.
a. Germanic Letters as Symbols.
Spirants.
Av - ^ t,, 6, te
h> j> A *.-
All scientific work in Avestan implies extensive com-
parison with Sanskrit ; on this account the Av. translitera-
1 Strictly the L should be varied somewhat from the 'tag' below re-
ferred to, in order to show that it represents a somewhat different modi-
fication.
H
tion is always to be brought into closest symmetry with
the Skt. transcription. This is practical and it is necessary.
It is necessary, however, on the other hand to avoid con-
fusion with the Sanskrit. Confusion sometimes arises from
using a symbol in Av. with a different value from that
which it familiarly has in Skt., or the same sign for sounds
that differ enough in Skt. and Av. to require a distinction
to be made between them. A striking point, for instance,
in which differentiation between the two may be made is
in the matter of the Av. spirants. The possession of
spirants as contrasted with the Skt. aspirates is one of
the characteristic phonetic features of Avestan. For scien-
tific purposes, then, these Av. spirants should be distin-
guished from the Skt. kit, gh, th, dh. Nor is this to be
done alone of account of the difference of sound — the
symbol having a spirant value in transcribing the one
language and an aspirate value in transliterating the other
— but also, it might be added , because confusion in Av.
may sometimes thus arise from the fact that gh in this
way should have to stand for the single character 9 Q)
and for the double letters <yc (gh) found in G(Y)Av., e. g.
GAv. jo-^xy-c^g^ aojon-g-Ji-v-af, fow&^h fsJn-gh-im, YAv.
■GHwce^gjro pjsocinghpm.
To obviate this possibility of confusion, recourse has
been had on the one hand to the Gk. characters /, y, 6, & (6)
by some (especially Hiibschmann, Umschreibiing der iran.
Sprachen; cf. also K.Z. xxiv. p. 323), on the other hand
to letters derived from Tuetonic signs ty, j, p, d (Rask,
Westergaard; esp. Pischel, Bartholomae). The general ten-
dency (e.g. cf. Brugmann, Grundriss ; etc.) seems now to
accept the Germanic rather than the Greek signs ; they are
therefore here adopted. Comparisons with Greek words,
moreover, are more common than with Germanic; less con-
fusion therefore thus results and such un-Avestan looking
words as ypaQpa are likewise avoided. In America, Eng-
15
land, and Germany the A. S. ]>, d, j if not actually on
hand are always easy to obtain whenever scientific tran-
scription is needed ; and ^ can be supplied by an Old Eng-
lish or German long h (fy). Almost all philological type-
fonts contain the former two, at any rate, of these signs.
They possibly are open to some objection from the Romance
side; but the substitutions offered under each should be
considered. These spirants may now be taken up in detail.
6. Av. b> // (Justi kh).
Av. *oyf<>*tyy //rateus (Justi khrate'ns).
Palaeographically the A v. character ly (kh) is derived
by the upward 'derivation stroke' from the Pahlavi -» h.
A good transliteration % was suggested for it by Pischel
(B.B. vi. p. 275). The character 7/ is somewhat similarly
derived from the long Gothic h. It may always be given
by setting an ordinary Old English or German long h (fy).
7- Av. ^j (Justi gh).
Av. -fo\> ujram (Justi ughreni).
For the spirant ^the symbol/ — the ordinary roughened
g. of the Anglo-Saxon — is used, as often. It is on hand
generally in philological type-fonts. If not, it can possibly
be given by the long j of 03 , or better the compositor
may set a simple Old English or German g (9 g g).
8. Av. 6 p (Justi th).
Av. -ufo-i*^ ]?rdta (Justi thratdy
The dental spirant 6 is derived from the Av. sign for t
by the upward derivation stroke. It is represented, as
often, by the common A. S. 'thorn' ; this is certainly found
in all type-fonts that do linguistic work.
i6
9. Av. ^ ^ (Justi dh)}
Av. -"^ zVfrz (Justi idhci).
The symbol <£ or </<J is the common A. S. ( edh' or
crossed d. 2
b. Diacritical Marks on Letters.
In transcribing Oriental alphabets there are always a
certain number of characters that may be well represented
by merely adding to similar Roman letters a diacritical
sign as mark of differentiation. It is proper that the
diacritical mark should be used consistently throughout.
Furthermore, since comparisons between Avestan and San-
skrit must constantly be made, it is necessary again to
avoid using a diacritically marked letter with a special
value in Sanskrit and then employing the same marked
letter is a slightly different function in Avestan. Confusion
arises from disregard of this point. Thus, Justi's / (jo) is
confusing as it represents a different sound in Av. from
that which / (cerebral, lingual) familiarly has in Sanskrit.
Again, h for Av. ^ would be confusing on account of
Skt. visarga. 3
A palaeograpjiic study of the MSS. shows that many
of the Avestan characters are actually derived from one
another or differentiated by a stroke I » j or curve which
Professor Geldner orally termed a 'derivation stroke'. 4
1 For the voiced labial spirant QW (Justi w) — derived from Qj by the
derivation stroke — the simple letter w with the German pronunciation, has
been retained as is generally done. A crossed b i. e. t> which is used by
some scholars in Gothic transcription , would perhaps well represent it
orthographically, but here strict conservativism seems preferable.
2 On } w, tf y, p" h see below.
3 Scientifically, however, visarga in Skt. seems better transliterated
by h or the like , the subscript point ( % ) being reserved for the Unguals
{, (h, d, dh, V .
4 For many interesting particulars connected with the MSS. in this
respect especially, and for valuable hints I am deeply indebted to the
kindness of Prof. K. F. Geldner.
*7
Thus \> h, to gh or 6-, <y o) ir, ><> £, -o i£ etc. — the dotted
line denoting 'derivation stroke'. 1 Systematic treatment
suggests that we should adopt some similar method to
differentiate the Latin letters whenever it is necessary to
designate such modification in transcribing Avestan charac-
ters. At the excellent suggestion of Dr. Geldner, a 'tag' (J
has been adopted to be used somewhat in conformity with
the 'derivation stroke' whenever it is necessary and possible
thus to differentiate. Practical grounds favor such adoption:
when the types are cast the 'tag' breaks oft less easily;
furthermore, it may always be mechanically constructed by
an inverted spiritus lenis ( c ) set close or even by the turned
apostrophe (J. The plan, therefore, is thoroughly practical;
the tag will also serve somewhat in recalling the palaeo-
graphic character of the letter it is used to designate. 2
To prove that the tag as modification can be readily em-
ployed, reference may be made to some Avesta Notes by
the present writer in the American Journal of Philology
1889—90.
The letters diacritically marked by the 'tag' either as
sign of differentiation or as a representative of the Av.
'derivation stroke' may now be taken up in detail.
10. Av. jo / (Justi /).
Av. jojAmj bar at (Justi bar at).
The Av. letter ^ is one whose phonetic nature is un-
certain. Justi employed / to transcribe it. This is open
to objection; the subscript point (.) brings in confusion
with the Skt. lingual /. Various other devices have been
used to represent it — see Appendix p. 31. Whatever the
1 See also Rask, Echtheit p. 57, 50, 'Zug', ' Aspirationszug' ; Spiegel,
Gram. p. 17 'durch Anfiigung eines Striches' ; Hubschmann , K.Z. xxiv.
p. 339 etc.
2 A 'tag' as modification sign is elsewhere in use, cf., for instance,
Sweet's Anglo-Saxon Reader ; Sievers-Cook's Grammar of Old English, et al.
2
1 8
phonetic value of the letter may have been, its palaeo-
graphic character is clear; ^ is formed by the 'derivation
stroke' directly from ro t. It may therefore be consistently
represented by t-\- £ , thus:
Observe that / is mechanically to be made as before by
means of the inverted comma shaved off and set close,
or by the turned apostrophe.
Nasals.
Av. } , -u, i, ^.
v, ty, n, 7%.
ii. Av. \ 10 (Justi n).
Av. g^oyi-" awhaf (Justi anhaf).
For the guttural nasal ? (Justi n) the symbol p derived
from the old long tailed Germanic n has been adopted.
This is the common designation for the guttural nasal in
linguistic works. 1 In short articles where scientific trans-
literation is required, the symbol P may be produced
mechanically by inverting a (v) and opening the bottom
with a penknife, thus v.
12. Av. -i/ ig (Justi n).
Av. t))j<y-tf-»^3 daitylwus (Justi dainhe'us).
The character -a is palaeographically a modification of r,
it occurs for \ v in connection with h when preceded by
an i- or ^-sound. The two ? -a interchange often enough
in MS. variants to show the resemblance that -a must have
had to ) in form and in force. The palaeographic cha-
1 The sign n (Justi) should be reserved for the Skt. palatal n
(Whitney «)— see the suggestion by Pischel, Gott. gel. Anz., 14. Juni 1882,
p. 738 seq. The letter v is preferable also for the guttural nasal in Skt.
(Whitney ft) when comparisons with Av. are made. y .*'
19
racter of -u, the transcription consistently recognizes by
attaching the modification tag (J to v, thus forming ig}
13. Av. j» 11 (Justi it).
Av. *pj£i^3 barztiti (Justi barenti).
Palaeographically # is composite character represent-
ing an tf-sound with n. In appearance it resembles x' 4 of
the MSS. Jn usage, however, ^ is merely a modification
of 1 n before stopped consonants. According to the system,
therefore, it will be represented by n with the modification
tag. This tagged n (n) recalls on the one hand the true
;/-character of ^; on the other its external resemblance to
the nasalized x* q,.
Sibilants.
Av. u, t), ^, yo — QU.
s > s > $> § — z -
14. Av. 13 s (Justi c).
Av. -vou-" asti, ^(yy-" aspd (Justi agti, agpo).
The letter « shares the nature both of the palatal and
of the dental sibilant. It is now universally given by s;
this is therefore adopted. 2
15. Av. -o ( e , n)) s (f, $), Justi s (sh, sk).
Av. -ojro-^ ratus. j vy^> i$us, ■^\6 / ^ M * j yv §yaopna (Justi ratns,
ishus, skyaothna).
From a palaeographic standpoint it is evident at a
glance that -0 forms the basis of the three ^//-sounds. The
sign s must therefore likewise form the basis of the three
transliterations. It is customary thus to transcribe -o by s
rather than by $, in order to avoid confusion with the Skt.
1 Justi's ri is open to the objection that scientifically the acute (')
should be reserved for designating accent.
. - For the palatal -s in Skt. (Whitney 4) it seems preferable to use s
when comparisons between Av. and Skt. are to be made.
20
lingual s noted above. Justi's s for -v has had of course
to be abandoned. Now since ^ is a composite character
made by the derivation stroke, it may be consistently
rendered by use of the modification tag, thus
& (U + -o) = $ C e + &
In like manner yo is palaeographically a modification of -o
before y. This differentiation by means of the > turned in
the other direction may be indicated by the reversed tag §.
The threefold differentiation of s, $, / is n o tjnecessary
except in transcribing a text for scientific purposes where
it is desired to reproduce the differences of the original.
For practical purposes, the sharp distinction may be quite
disregarded ; -o is final (except before t, c) , ^ initial and
internal, yv only before -». The simple s will therefore
practically suffice i. e.
or all simply s.
The sign i is on hand in all linguistic type-fonts. The
differentiation if found necessary, may of course easily
be made as before by means of the subscript tag. 1 This
concludes the sibilants. 2
Aspiration.
Av. or-, £, \».
h, k, hr.
The fundamental aspirate is <y h (Justi h). It is a
modification of the Pahlavi - h. In the transliteration of
it, all are agreed. In regard to the other two ^ ^ there is
a diversity in treatment. Justi wrote both as q; but they
1 The ligature t&o ft, KK) Ic require no further special designation.
2 The voiced spirant sibilant <tt) (Justi z/i) is commonly rendered i
in conformity with /. No remark is needed.
21
are sharply to be distinguished. They will be examined
in detail.
1 6. Av. £ k (Justi q).
Av. -6k*i )«t"^3 dafyyunqm (Justi daqyundm).
The character ^ is a mere modification of h (Pahlavi -")
before a y. It may be thus represented :
$ (^ + -) = ^ ^ + ^;.
The //-sign (Pahlavi -«) is the basis ; the 'derivation stroke'
is represented by the 'tag'.
17. Av. v h (Justi q).
Av. -"r> h/a- (Justi £#-).
The letter ^ is a ligature of Pahlavi -" h and 1 v. It
interchanges at times with »<y hv. To transcribe this liga-
ture, the suggestion is here made that we may aptly use
hr, in itself a similar ligature, that is often used in Gothic
transliteration, e. g. Braune's Goth. Gram, transl. by Balg,
1883, cf. § 63. Thus:
|» (| JL») =s fff (k V).
As this sign is not always on hand, its place may be
supplied by h with a slightly raised v, thus kv.
Semivowels.
Av. ro («) y (Justi y), ^ (») v (Justi v).
Av. >"mroro yesnyo, ^^)^ vidvw (Justi yepiyo, vidhvdo).
The characters ro £,), as is familiarly known, occur when
initial; the forms *> » are found when internal. The simple
transcription y, v for both the initial and the internal forms
has been retained on conservative grounds. The same are
retained likewise by Brugmann, Grundriss der indogerm.
Sprachen. Confusion can hardly arise; the exceptions to
22
the law of the initial-internal forms need scarcely be
noticed. Scientifically, however, it is more accurate to
adopt i u for » » as has been done by others. Thus :
™ y \ , , ,
or both simply y.
" k \
b V I
or both simply v.
» u \ * J
Palatals.
Av. P c (Justi c), i j (Justi j\
A v. -&M j h ciprzm, »to**we jasaiti (Justi clthrem , jacaiti).
The voiced palatal ^ is palaeographically a derivative
from v. In transcribing y ^ it seems best to be conserva-
tive; Justi's c, J are retained; they are likewise kept by
Brugmann, Grimdriss. The same is almost universally the
case in transcribing Sanskrit. Those who wish to be more
scientific in this respect are welcome of course to the
palatal point over k, g, thus k, g.
Resume.
Such in the main are the characteristics, palaeographic
and phonetic, of the letters in the Avestan alphabet upon
which there is most discussion ; such likewise is the system
of transliteration proposed. I have adopted it for my
Avesta Series: i. Grammar, ii. Texts, shortly to appear.
Many of the points in regard to the alphabet are, to be
sure, more or less familiar ; in such cases the transcription
adopted has merely followed what it seems the general
tendency to adopt. May these points become more and
more universally agreed upon! The new points in the
transliteration are suggestions toward uniformity or toward
improvement by remodelling. On the latter, compare for
instance the discussion of the ^--vowels.
23
The principal features of the transcription suggested
may now, therefore, be recalled. They are the consistent
use of the 'tag' as a diacritical mark, instead of points
or accents. This 'tag', let it be remembered, answers in
general to the 'derivation stroke' by which palaeographi-
cally so many of the Av. characters are formed or modified.
Furthermore, the system marks clearly the orthographic
distinction of the three sibilants i, f, $, when necessary,
also of the nasals p, y, n, n, and of the aspirate h, J{, h.
A practical transcription of the much-discussed % / (Justi /)
is suggested. An innovation is made by remodelling the
^-vowels, using b, 9 as an apt representation of the un-
certain ? {, and thus bringing k> e ^ e 3 ^» o > 5 into closer
relationship — however the latter be rendered. With re-
ference' to y c, ^j, w w, ro » y, 1} » v, conservativism
has been used. The Germanic characters Jf } d, etc. have
been adopted for the spirants according to what seems
to be the apparent tendency of the present.
In conclusion I would like once more to renew my
cordial thanks to each of the scholars mentioned at the
beginning of the paper. Their courtesy, their suggestions
and their advice are cordially appreciated. To Professor
Geldner, as always, is my gratitude due for the interest
that he personally took in discussing the various sides of
the question of Avestan sounds and of their transcription,
as well as for the trouble he went to, not alone in giving
me valuable points in regard to palaeography, but also
for making accurate copies of many letters and styles of
MS. writing. To these obligations may be added my in-
debtedness to the publisher, Herrn W. Kohlhammer, and
to the compositor, Herrn Sauberlich, for the form and for
the accuracy with which the paper is presented.
It remains alone to repeat the hope that Avestan
scholars may strive more toward union in transliteration.
24
The above system represents in most points what seems
to be the general drift in regard to scientific transcription
of the Avesta; it has only added or modified where altera-
tions seemed necessary; and the practical side of the
question , withal , has been kept as much as possible in
view. If its main points should receive the sanction only
of certain scholars, how many there would be would
follow these!
A. V. Williams Jackson
July 1890. Columbia College
New York City.
Appendices.
Appendix I.
Proposed Transcription of Avestan.
(Compared with Justi, Handbuch der Zendsprache)}
A. Voxels.
Short M a * i i * ? ' ? \u e ^> d
a i u (e) (e) o
Long *» d t* t * u \d ^e > r» m j? q
a i u (e) S o (do) (a)
B. Consonants.
Guttural 9 k b fy C g i, J
k (kh) g (gh)
Palatal \> C — % j —
c J
Dental r» t p _J d ^d £ /
/ (th) d (dh) (()
Labial <y p <& / . J b w IV /
p f b w
Nasal i *> iS ty I « #U 4 M
(h) (ri) n (n) m
Semivowel and
Liquid ro (*») y (i) 2 ^ r tj (») s> r«; 2
y r V
Sibilant 13 S *v S g}/ S TO $ fB OJO^
(0 (s) (sh) (sk) z (zh)
Aspiration . . . . <y k ^ k
h (q)
Ligature \» h r
(q)
1 Forms in parentheses ( ) show where Justi has been deviated from.
2 The signs i, u need only be employed for purely scientific pur-
poses; the letters j, v for both initial and internal ro *», lp », answer
fully for practical purposes.
3 The differentiation /, $, / need only be made in scientific articals.
The single sign s is ordinarily quite sufficient for the three "O, £0, ro.
Appendix II.
Suggestions to the Printer.
For small articles when special types not cast.
(See Alphabet, Appendix I.)
Vowels 9, 9, M.
The p is simply a 'turned' e. — $, turned j with macron
(") as quantity mark. — For co, set a a close together by
shaving edge of type a little ; then unite under macron.
Unusual Signs ty, J, p, d, 10, Jr.
For fy, j } set Old English or German or similar long
It, g. — p, d are Anglo-Saxon 'thorn', 'edh'. — w, invert a (v)
open and shape with penknife v, or set different ;/. — h,
substitute h v .
Tagged Letters /, #, n, }{, q.
Use a turned spiritus lenis, or set inverted apostrophe.
— q, differentiate tag somewhat, if desired.
Optional Letters j t $, i, u.
Presumably, s is on hand. For $, /, if needed in
purely scientific articles, set tag and cedilla beneath s. —
For /, u, if likewise needed, shave off ^ and set close
beneath.
i
Appendix III.
Substitute Alphabet. 1
For popular articles.
(Modelled after Justi.)
A. Vowels.
Short a i U B O
Long a I u p o m q [or a\
B. Consonants.
Guttural k kh g gh
Palatal c — j —
Dental t th d dh /
Labial p f b ZV
Nasal . . . & [or n] % [or ft] n fi [or n] m
Semivowel and Liquid y r V
Sibilant s S [or sk] Z Z [or z/l]
Aspirate Ji fy
Ligature k» [or q]
1 For suggestions to the Printer see Appendix II.
Appendix IV.
Some of the different Systems of Transcription.
(References to the following Works — see Bibliography below.)
Anquetil du Perron — Zend Avesta (1771) — Vol. ii. p. 424.
Rask — Alter u. Echtheit d. Zendsprache, iibersetz. Hagen
(1826) pp. 46, 81.
Burnouf — Le Yacna (1833) p. xxxvii seq.
Spiegel — (Works) e.g. Av. iibers. (1852 — 63); Commentar
(1864—8); VergL Gram. (1882).
Haug — Gathas (1858 — 60); Essays (1862 — 84); Zd. Pahl.
Gloss. (1867).
Lepsius — Das urspriingliclie Zendalphdbet (1862).
Justi — Handbuch der Zendsprache (1864).
Roth — (Contributions) e. g. Kalender (Z.D.M.G. xxiv); Bei-
trdge (Z.D.M.G. xxv); Yacna 31 (1876).
Fr. Muller — Zendstudien iv (Wien. Ac. Mai 1877); Beitrage
u. Mittheilungen (Wiener Zeitschrift 1887 seq.).
Hubschmann — Iran. Studien (K.Z. xxiv. p. 370. — 1877 — 9);
Umschreibung (1882).
Pischel — Umschr eibung (B.B. vi. p. 272. — 1881); Recension
(Gott. gel. Anz. 1882 p. 737).
de Harlez — Manuel de la langue de r Avesta (1882).
Sacred Books of the East — Translit. of Oriental Alph.
Geldner— Metrik (1 877) ; Studien (1 882) ; Drei Yasht (1 884) ;
(later contributions) in K.Z. and B.B.
Bartholomae — Verbum (1878); Gathas (1879); Ar. Forsch. i.
(1882); Handbuch (1883); A. F. ii— iii. (1886—7),
and (later contributions) in K.Z. and B.B.
Brugmann — Grundriss der vergl. Gram. (1886 — English
transl. 1888).
30
Systems of
(Parentheses denote forms
jj
J-W
y
y
>
•p
I
S
K)
«
-s
>
f*»
)C
Proposed
Transcription
a
a
I
t
u
u
3
9
e
e
CO
%
Anquetil du Perron
a,e a
e
i
021
d
/
i
e
do
an
Rask
a
a
I
71
U
ce[ci\
e
a?
e
e
do
%
Burnouf
a
a
i
i
21
U
e
e
e
3
do
a
Spiegel
a
a
i
i
U
u
(e)e
e
(e)e
(to
do
(an)d
Haug .
a
(7
7
i
7/
a
e
e
e
e
do
a
Lepsius
a
d
i
l
?/
u
e
e
e
e
o
a
a
Justi
a
d
i
7
11
u
e
j
e
c
do
a
Roth
a
d
i
7
1/
u
e
4
e
e
do
a
Fr. Muller
a
a
i
i
U
u
»/
w/
e
e
o,
o
d
a
Hubschmann ....
a
a
i
i
u
ii
e
e
e
e
d
(a)q
Pischel
a
a
i
l
u
11
e
e
e
e
o
a
4
de Harlez
a
a
i
I
u
u
e
J
e
e
d)
a
Sacred Books ....
a
a
i
i
u
u
e
e
e
e
o
6
au
a
Geldner
a
_ .
a j i
z \u
u
e
w
(e)e
e
do
a
Bartholomae ....
a
a
i
Z 11
ft
e
e
e
e
(do)d
(d)q
Brugmann
a
a
i
I
u
it
e
e
e
e
o i d
a
Proposed
Transcription
a
a
i
I
u
u
$
d
e
e
(V
4
3i
Transcription.
temporarily used by an Author.)
9
b
c*
i
V
i
r=>
6
J
t,
5
k
i
g
J
c
J
t
J>
d
d
/
k,C
kh
g
gh
tch
dj
t
th
d
d
d
k
2
g
\
c
j
t
P
d
d
i
k
kh
g
gh
tch
dj
t
th
d
dh
t
k
kh
g
gh
(ch) c
j
t
th
d
dh
{0 4
k
kh
g
gh
(cli) c
ii)j
t
th
d
dh
(t, t) 4
k
m L
g
[£\ T
{k)c
[g\j
t
[/] 6
d
<?
8
k
kh
g
gh
c
j
t
th
d
dh
t
k
kh
g
gh
c
g
t
th
d
dh
(ft f) t
k
L
g
T
c
g
t
e
d
I
t
k
(/J *
g
T
(c)c
U) J
t
6
d
I
&§ti
k
*
g
J
(c)k
(J)g
t
t
d
d
i> i
k
k
g
g
c
j
t
t'
d
d'
t
k
kh
S
gh
k
g
t
th
d
dh
t
k
k
kh
(kh r/j x)}i
g
g
gh
c
(k, c) k
(g)j
(gj) g
t
t
th
d
d
dh
(dh,l)d
(0/
k
X
g
T
c
j
t
}
d
d
J> i
k
t
g
J
c
j
t
P
d
d
t
32
Systems of
(Parentheses denote forms
Proposed
Transcription
Q) &
p f
b w
TO
y
y
Anquetil du Perron
Rask
Burnouf
Spiegel
Haug
Lepsius
Justi
Roth
Fr. Muller
Hubschmann . . . .
Pischel .
de Harlez
Sacred Books . . . .
Geldner
Bartholomae . . . .
Brugmann ......
Proposed
Transcription
P f
P f
P f
P f
P f
p[P'¥
P f
P f
p f
p f
p f
P f
p f
p f
P (?)/
p f
P f
b v
b v
b w
b w
b w
b [b ( ]v
b w
b w
b (w)v
{p)w
b w
b ' w
b w
b iv
b (p)w
b w
b w
y
y
y
y
y
y
j
j
y
y
y
y
U)y
J{y)
y
i
y
y
y
y
y
y
j
j
y
Mi
y
y
U)y
Mi
y
V
IV
V
V
V
V
V
V
v)w
V
V
V
V
V
ou
w
V
V
V
w
V
V
(v)w
V
(v)u
V
V
V
[y)u
g
(n) 7t
it
it
n
n
n
in) w
n
ng
n
Iftj n) v
n
ng
i
{ii) n, n
n
n
n
n
(if) n
(it) n
n
N
n
n
V
33
Transcription.
temporarily
used by an Author.)
&
<>
i)
"O
gj
W
/
OJU
or
£
K"
-6
n
m
s
s
tf)i
Cfl *
~
z
h
h
fo
'fn
an
m
s
sch
sch
sch
z
j
h
kh
kh
km
N
to
Q
s
>
s
sk
z
■>
z
b
X
X
M
ft
to
C
s
ch
sk
z
j
h
q
q
hm
ft
m,
w
(s) sh
(sh) s
(sk) shk
z
(j')zh
h
q
q, hv
htn
ft
tn
«*
(s) sh
sh
(sk) shk
z
(z) zh
h
q
q
—
[d] ft
in
[i\s
s
s
sk [i]
l*\*
z
h
V
V
'm
it
to
C
s
sh
sk
z
zh
h
q
q
—
ft
to
f
s
sh
sk
z
zh
h
q
q
—
ft
to
s
s
sh
sk
z
z
h
q
q
—
ft
to
s
s
s
(sk, s) s
z
z
h
(#)*
(hf)xw
hm(m)
a
in
s
s
s
—
z
z
h
h-
(*)*
—
n
in
C s
s
sh
sk
z
zh
h
fo>
hv
hm
n
m
s
s
sh
sk
z
zh
h
hv
hv
—
(n) ft
to
to*
(s,s,sh)s
(sh)s
(sk, sh) s
z
(zh)z
h
(fl*
q hv
—
(ft, ft) n
to
s
(i s) *
(sh,s)s
(sk,s,s)s
z
(z,z)z
h
(q,h')h
(q,hv)h
tn
n
to
s
s
s
s
z
z
h
X
xw
—
n
to
s
s
to?
w *
z
z
h
h
hr
m
Appendix V.
Partial Bibliography.
Writings on the Avestan Alphabet and its Transcription.
Anquetil du Perron — Zend-Avesta, Ouvrage de Zoroastre,
ii. p, 425. — Paris 1 771 .
R. Rask — Ueber das Alter und Echtheit der Zend-Sprache,
iibersetzt von F. H. v. d. Hagen, pp. 46 — 81. — Berlin
1826.
E. Burnouf — Commentaire sur le Yafna; L' Alphabet Zend,
pp. xxxii — cliii. — Paris 1833.
H. Brockhaus — Vendidad Sade, p. xii. [Alphabet.] — Leipzig
1850.
— in Zeitschrift der deutschen morgenlandischen Ge-
sellschaft, xvii. p. 539.
C. Arendt — Phonetische Bemerkungen (Kuhn's Beitrage ii.
p. 429). — 1 861.
M. Haug — Essays on the Parsis (1 ed. p. 52 — 57; 3 ed.
West, p. x). — Bombay, London 1862 — 1884.
— Zand-Pahlavi Glossary p. 81. [Alphabet.] — Bombay,
London 1867.
Lepsius — Das urspriingliche Zendalphabet (in philolog. u.
histor. Abhandlungen d. kgl. Ak. der Wissenschaften
zu Berl. 31. Marz u. 14. Juli 1862). — Berlin 1863.
F. Justi — Handbiich der Zendsprache p. x seq. — Leipzig 1 864.
Fr. Spiegel — Gramm. der altbaktr. Sprache. — Leipzig 1867.
— Vergleichende Gramm. der alteranischen Sprachen. —
Leipzig 1882.
— Znr Geschichte des Avesta Alphabets (Bezzenberger's
Beitrage, ix. p. 173).— 1885.
35
Kuhn & Schleicher — Umschreibung des altindischen und
altbaktrischen Alphabets (Kuhn u. Schleicher's Bei-
trage z. vergl. Sprachforschung v. p. 144). — 1868.
Fr. Muller — Zendstudien iv: Ueber die Zischlaute des Alt-
backtrischen. (Sitzb. d. k. Akademie d. Wissensch.
—Mai 1877.) — Wien 1877.
H. Hubschmann — Iranische Studien i : Ueber den lautwerth
des Zendalphabet. — Mit 3 Tafeln. (Kuhn's Zeitschrift
xxiv. p. 323 — 427.) — Berlin 1878 — 9.
— Die Umschreibung der iranischen Sprachen und des
Armenischen. — Leipzig 1882.
C, Salemann— Ueber eine Parse?ihandschrift der kaiser/,
offentl. Bibliothek zu St. Petersburg. — Leyden 1879.
[Could not consult.]
C. de Harlez — De V Alphabet avestique et de sa Transcrip-
tion. — Paris 1880. [Could not consult.]
— De la Transcription de V Alphabet avestique (Bezzen-
berger's Beitrage vii. pp. 127 — 139). — 1883.
Sacred Books of the East — Transliteration of Oriental Al-
phabets (see end of each volume). — London 1880 seq.
R. Pischel — Die Umschreibung des Baktrischen (Bezzen-
berger's Beitrage iv. pp. 272 — 282). — Gottingen 1881.
— Recension von Bartholomae' s 'Arischen Forschungen i '
(Gotting. gel. Anz. 14. Juni 1882 p. 737 seq.).
Kirste — Die constitutionellen Verschiedenheiten der Ver-
schlusslaute im Indogermanischen p. 7 seq. — Graz 188 1 .
[Could not consult.]
K. F. Geldner — Metrik des jiingeren Avesta p. xiv. |Mere
mention.] — Tubingen 1877.
— Studien ztim Avesta i. p. 4. [Mere mention.] — Strass-
burg 1882.
— Miscellen aus dem Avesta. [On e, ».] (Kuhn's Zeit-
schrift xxvii. p. 257 seq.) — 1883.
— Drei Yasht aus dem Zendavesta pp. vi — xv. [On the
sibilants.]— Stuttgart 1884.
Chr. Bartholomae — Das altiranische Verbum p. ii. [Alpha-
bet.]— Mtinchen 1878.
— Die Gdthds: Metrum, Text u. s. w. p. 3.— Halle 1879.
— Arise he Forschungen i: Anhang pp. 155 — 163.— Halle
1879.
— Ar. Forsch. ii — iii Vorwort. — Halle 1886 — 7.
Beitrage zur altiranischen Graimnatik i. [On sibilants
and nasals.] (Bezzenberger's Beitrage vii. pp. 188 — 195.)
-1883.
— Handbuch der altiranischen Dialekte. — Leipzig 1883.
E. Dillon — Die Umschreibung der eranischen Sprachen
pp. 1 — 16. — Leipzig 1883.
D. Peshotan Sanjana — Civilization of the Eastern Iranians
Vol. ii. pp. 273 — 286 (The Iranian Alphabets, transl.
from Spiegel's Eranische Alterthumskunde iii. pp.
759—771). — London 1886.
P. de Lagarde — Mittheilungen ii. Bemerkungen iiber die
Awesta-Schrift pp. 38—48. — Gottingen 1887.
K. Brugmann — Grundriss der vergleichenden Grammatik
der indogerm. Sprachen Bd. i. p. vii, 25 ; Bd. ii. Vor-
wort pp. vii — viii. — Strassburg 1886 — 1889. — English
translation 1888.
RETURN CIRCULATION DEPARTMENT
TO— *• 202 Main Library
LOAN PERIOD 1
HOME USE
2
3
4
5
6
ALL BOOKS MAY BE RECALLED AFTER 7 DAYS
1 -month loans may be renewed by calling 642-3405
6-month loans may be recharged by bringing books to Circulation Desk
Renewals and recharges may be made 4 days prior to due date
DUE AS STAMPED BELOW
■ 11.11 2 71979
NOV 91982"
4
flCT22
'82
AUG01B97
13.
APR 16 20)0
fFB 7 2002
UNIVERSITY OF CALIFORNIA, BERKELEY
FORM NO. DD6, 60m, 1 1 /78 BERKELEY, CA 94720
®s
■BBtfTttu-
U. C. BERKELEY LIBRARIES
CD^bfilDSE
UNIVERSITY OF CALIFORNIA LIBRARY