BO UTAS

The Jewish-Persian Fragment from Dandan-Uiliq

For Stiy W ikanrler on his (Wth bi-rtlulny
rli
he ,Jewish-Persian (.JP) fragment from Dandan-Uiliq near Khotan
in East Turkestan was published for the first time in .JRAS 1903 (p.
747-758) by D. 8. MARGOI,IOU'rH together with an introductory note by
it.s discoverer STEIN (p. 735-746), additional notes by W. BACHER
(p. 758/7()()) and a rather imperfect facsimile (p. 7f)()/7()] ). On the basis
of this facsimile C. (K. Zaleman) published a careful revision
of the text in his 1trtielc ''Po povodu jevrcjsko-persidskago otryvka iz
Xotana" in the Zapiski Vostoenago Otdelenija rn (IH04-05, pub!. S.-
Petcrburg 1906), p. 04H-057. There he corrects many mistakes in
MAIWOLIOUTI-I's preparation of the text with special regard to Middle
P<'rsian (MP) forms and with application of a more Htrict New Persian
(NP) grammar. Unfortunately the indistinctJ quality of the Hmall facsimile
at. his diHpmial limit.eel his attempts.
MARGOLIOU'l'H's article on the fragment was republished with certain
ehangcs in Appendix C to Aun.EL STEIN'S detailed report on his journey
in IH00-01, "Ancient Khotan" (1907, vol. I, p. 570-574), and in the same
work (vol. 2, pl. 119) there is a very good facsimile of the fragment .
. Lat.er, important points have been elucidated by W. B. HENNING in
his article on the old ,JP inscriptionH of Tang-i Azao
1
and, more exten-
sivoly, in his contribution "Mittcliranisch" to Handbnch del' Orientalistik
I : 4: 1 ( 1958: p. 79-80). B.ecently the document has been reproduced
on the basiH of the two articles by MARGOLJOUTII in the joumal Hilal,
Karachi.
2
Although due rdernnce ha::i belHl made to this fragment in surveys
1
UHOAH 20 (I !J;ii), p. 33fi-342; see especially p. :l41, n. 2, :J, .t-, and p. 342,
II. I, 3.
2
IO (l!JU3/I:H2/I:J82): 4, p. 44-.')3: M. H.. u. KA.s(lAiti, l:lafl)ah'i az giisah-yi
foraml'18-sudah-yi tarix, xadnmiit-i Turkistan bi-;1,abiin-i fii1si va kaHf-i qndimtarin
yiiddailt.-i fiin;i az X utan.
26 From Old to New Persian
of JP and old NP sources,
1
it has perhaps not attracted the attention
it deserves, being one of the oldest, if not the oldest example of NP
language preserved in a document of some length. MARGOLIOUTH's
original dating, about 100 A.H./718 A.D. (JRAS 1903, p. 748), is cer.-
tainly not acceptable,
2
but all available evidence points to the second
half of the 8t.h century A.D. as the most probable origin of the document.
This evidence is, in short, the following:
1. We must accept the judgement passed by AuREL STEIN on the
circumstances of the discovery of the fragmcnt
3
and presume its pro-
venance to he from one of the ruins of the desert site called Dandan
Uiliq (or Dandan oilik, 'the Ivory house').
2. This Dandan- Uiliq, situated some seventy miles to the northeast
of the Khotan oasis, was known to the Chinese of the T'ang period as
Li-hsieh
4
and had in the second half of the 8th century a convent of
Chinese Buddhist monks (STEIN 1907, 1, p. 275 ff.) and a minor imperial
garrison with a Chinese commandant (SrEIN 1907, 1, p. 266 f.); it was
obviously not situated on the direct route of the southern branch of the
Silk Ro.ad but somewhat to the north of it.
5
3. Archaeological evidence indicates that Li-hsieh was finally deserted
about the end of the 8th century, probably in connection with the col-
lapse of Chinese authority and the Tibetan invasion (STEIN 1907, 1,
p. 283 f.), and this chronology is confirmed by geological investigations
which elucidate the desiccation of the Tarim basin in general and the
gradual recess of the arable fringe east of Khotan in particular.
6
1
E.g. M1NORSKY, JRAS 1942, p. 183; FISCHEI., The Jews, their history, culture,
and religion, ed. by L. FINKELSTEIN, 1949, p. 823; JAOS 85 (1965), p. 151; HENNING,
Akten des 24. Internationalen Orientalisten-Kongresses, 1957 (publ. 1959), p.
305-7; Majallnh-yi Dii.niskadah-yi Adabiyii.t, Tehran 1337/1958, p. 2; OnANSKIJ,
Vvedonie v iranskuju filologiju, 1960, p. 248; LAZARD, La langue des plus anciens
monuments de la pl'Ose persane, 1, 1963, p. 31; ASMUSSEN, Acta Orientalia 29
( 1965), p. 49 f.
2
Cf. already SALEMANN, op. cit., p. 053.
3
JH.AS 1903, p. 737 ff.; Ancient Khotan, 1, p. 306 ff.
4
Or Li-sieh, Li-sie; cf. CHAVANNES in STEIN 1907, 1, p. 521 n. I.
5
Cf. HERRlllANN, Die nlten Seidenstrassen zwischen China und Syrien, 1910, map.
8
Cf. STEIN 1907, I, p. 95 f., 285 ff., 306, 369 ff., 42!), 461; Dandii.n-Uiliq/Li-
hsieh was obviously once the "terminal oasis" of the Chira river or one of its neigh-
bouring streams and not, as argued by HEDIN (Through Asia, 1898, p. 798; Die
geographisch-wisimnseht1ft.lichen Ergebnisse meine1 Reisen in Zcntralasien, 1894-97,
1900, p. 37) and H E R R M N ~ (HllO, p. 97 f. and map), watered by the more eastem
Keriya-darya which would have had to follow a different course; on the reduction
The Jewish-Persian Fragment from Dandan-Uiliq
27
4. An analysis of the paper of the fragment by J. WIESNER (JRAS
1903, p. 742 f.) has shown it to be old and not distinguishable from the
paper of the Chinese documents found by AuREL STEIN at Dandan-
Uiliq, some of which carry dates corresponding to 781, 782, 787, 789,
790 A.D. (CHAVANNES in S'1.
1
EIN 1907, 1, p. 533).
5. Palaeographically the fragment is quite archaic, belonging to a
definitely older stage than the JP "Law report from Ahvaz" of 1020
A.D.,
1
and resembling rather the inscriptions of Tang-i Azao from 752/3
A.D.,
2
and this archaic character is confirmed by linguistic evidence,
as will be seen below.
The document quite clearly consists of a fragment of a business letter
concerned mainly with the trading of sheep and possibly also cloth(ing)
and slaves. So much of the text is missing at the beginning and end of
each of the 38 lines that a clear syntactical connection from one line to
the other is nowhere to be established. On the other hand nothing seems
to be lacking either from the top or from the bottom of the sheet, and
there is even a possibility that lines 1 and 38 form the first and last of the
letter.
The orthographic system agrees in its main features with the one used
in other JP texts. As for the consonants, it should be noticed that no
diacritic signs are used, that i' represents k and :::> x (as in the inscrip-
tions of Tang-i Azao), i1 for his written in a closed form not distinguish-
able from n, :!!: is used for j as well as c, D for p as well as /, and :::i, l, i
do not distinguish between plosive and fricative pronunciation of b, g,
d / {J, y, <5.
3
In a few cases it seems as if the initial vowel could be written
without K, but possibly these instances could be explained as enclitic
constructions (lines 22, 24, 31). Vowels are represented according to a
fairly fixed scheme: ii - K or defective scriptum, i/e - ", u/o - i (at times
"and i nearly coincide in the actual writing); a is generally not represented
(although final a with K), i is not represented or written out plene "
of the watcrflow in the rivers and its probable dependence upon a shrinkage of the
glaciers of the border ranges (here K'un-lun), see E. NORIN, The Tarim basin and
its border regions ( = Regionale Geologie der Ertle, 2: IV: b), 1941, p. 30 f., and on
possible connections between the size of glaciers, climatic changes and historical
processes, see B. ALMGREN, Bull. of the Mus. of Far Eastern antiquities, Stockholm,
34 (1962), p. 97 f., 101.
1
Cf. A. E. COWLEY, JRAS 1903, p. 750 f.
2
Cf. w. B. HENNING, BSOAS 20 (1957), p. 339.
3
Cf. D. N. MACKENZIE, BSOAS 31 (1968). p. 249 f.
28 From Old to New Persian
(especially as connective vowel l><'forc enclitic elements), and u is written
plene i (when not epenthetic).
The main object of the present article is to facilitate future studies
by providing a serviceable photograph of the fragment, now kept in the
British .Museum (see p. 127), together with a suggested reading of the
kxt in Hebrew letters (partly ascertained from the original), a transcrip-
tion1 and a t.cntative translation (word for word). The treatment of the
text leaves many problems unsolved and the work should be regarded
mainly as a presentation of the material.
In t.Iw t.cxt, transcription and translation the following symbols have
hPcn used:
( ) not clearly readable
I ] possible addition or (in the translation) added for clarity
! i (in the translation) to be ignored
/ alt.ernati V<'l.Y
deletion
1
I vcnt.u1c to give a voealised trans<"ription instead of the mere transliteration
of the eonsonant frame currnnt in JP studies, as I judge a full 1eprmmntation of the
words to be easier to survey in the fragmental'y state of the present document; it
should be notic-ed, howovel', thnt this transcription is r.om1rntional n.ncl not based on
a full phonnmi< or analysis.
The Jewish-Persian Fragment from Dandan-Uiliq 29
W-:. 0 1" l ~ I ~ ( Ui(i ) .
.1'11hlis l1t d h.1 p1r111 issio11 of the Trn, k eti of t he llritisli ,\11 i;;p1
11
n.
30 From Old to New Persian
Text
l)K T11 11T 1'!VK:I 1K" "K i111::1 1"T" ( 1
)'IV ":I Cl111j:> '!V":I nO":l1 i1(1 /i 2
)nT1 KE>1 101 "l KE> "K ip 1(ll 3
):I "lN 11N1 ":I 'IV" 101 "0 N( 4
'IV) iiiE> i:iN Nn .,,::i 1"1:1"( 5
)j:>"(l) "K '!V"1E> mi::i 1"T" Kiin wi::i "K l< 6
N) i11 Nm "(N) i11l Nn 1NnE>i ":I iii (1/1 7
i)::>"::> mi::i 1"T"i ili::> inoioi ii:i ":t 1lE>O( 8
)lN'!V" 'J'll 11:1 Nl (1)10(1>) lK'IV" TK op "l 1Kl1j:>( 9
):IN "11 NE> l("K n)ON11:1 i1n::l11E> "K ("l 10
l/N)1E> 1i1'!V "N (10) 11:1 Nl op 11:1 i1n::l11( 11
i/") nOi1 i11"E> Nl 1(1l) T"l1 CIN1::> "N ipnE> i1"0 l(K 12
1)Np "P" ciiNi in NE> "lN 11l ( 13
) "'":t'!V ln=>Nl'!V ":I "N lN"T1 iio( 14
l) Nn 11"1::1 "N "10 TN ilE>01l qil 15
) i1"0 ":1(1) ip "nE>1l 111"N( 1()
)Oi1 lN"T n::ioi 17
rpin T" "N "iO TN::: N11 (N 18
)li ii::ii "n::>iiE> ii::ii .,,.,,::> ii::i ip< 19
) lii:i "nO"N:I n::i:i iio ilN q 20"
)1) milN T"l '!Vi:> "N= = TN in ":I (" 21
n>oiii 1,,., ilE>Oil "N iioi ii:i ,,.,N "l1 .. ,.,n0"1( 22
)1!JN:JO p<"i)Tl ln1 ii1N1::1 , .. T" lil ioi in< 23
) "i'" ip i<"E>)1j:> 1,,., i:iN:JO N1 "( 24
moi :ii: illi Ni pr"lP "1N 25
):It "P" p:iiil (T)N "!l ciiNi(l"):I Nl ciil":l i1!l(N 26
)"O "N p"iilN "!l Ni .,,.,1:i Nn cii1i i<::i/:i 27
)N , .. nE>il lN TN i1":l "P" ":l cinE>" ip ('IV 28
) ii:i mip Nn cuip n::>C( )'!VN "N iNp (N 29
"/i>:i i1noi::i Ni ci'!Vii1 "l ii::i milN T"l ciwii1< :10
")nE>ip lii" cii"oiiE> i1n:iii nwic1=> "N "io TN< :11
p>i 1(0) "1NnON1::1 "10 lN TN ii::ii lN11E> NE>( 32
):!tlE>i io NE> ip .. , .. n0"1!) iilN( 33
"N l(N) ilE>Oil (")N T"'!V!) JN "N i11l(i::> 34
)N "iO TN "(N) i1i?;lNl Jii":i ii1(W 35
) ?Niii l"!l"p1i l"T TN W 36
TN) 1"1!:1 "N T"i1 "l 1i1 TN ( 37
ci::i 38
The Jewish-Persian Fragment from Dandan-Uiliq
Transcription
) xuaoah i yar basao ziio roz '(g
2 r/d)h u-bist namah bes kardum be s([uma '1]
3 niga)r kii jamak? i man pa ce rasao u-pa dzt(?
4 )' sc rasao-is be farmai daoan ane mara b(
5 ),11bryn? xare ta man abar xastume furoo suoume (8
() )n i xua;; tora yazeo xuaoah pao-is muzd i (n)ck 'I (
7 r/d) diir be-uftao ta nuh i? mah u-ta dah (i?
8 fgoJ)spand be biio u-sust[t]ar xarand u-yazeo xuaoah
xyx(w?
9 )kunao cc kas az-isan (far)sfi(o)'? na buvao ciin-isan(
10 ce?) jamah i furoxtah biio-as(t i)n? pa roy-ima 'b(
II [fu])roxtah biio kas na bwr! sao? marduman i sahr pd('/g
12 a)n? sih *pitkii e xaram u-<"iz (nar)d'? na-paidah hast (y/\\'
rn ) (din ane man pa to u-m(ar<l) <larum yakc ka(r- '?
14 )soo u-z.iyiin i man be sinaxtan u-9 *8abili? (
15 Cim) gospand az soy i man-ra xarioan ta (g
rn m )eoiin gufte kii (ra)bbi'? sih (
17 l_aJ)maoan u-saxt ziyan-iimand has([t-'IJ
18 ') vara az soy i man 17 to(p?
19 )kit xuao xarioe u-xuao furoxte u-xuao ran([jiol''!]
20 n) agar man-ra soo baxt bayiste bf1oan man (
21 y) be to az sumar i xuas ciz andoh ma far([mai]
22 r f i]) restioe u-ne eo ar b iio u-soo i gospan<l-i8ftn '? du ru s ( t
23 )to rasao ciin yazeo xuahao u-tan-isuma naz(di)k sb'bd?(
24 )y-ra sb'bd?-ioftn? kw(py)d? kii mara *camxiii yake (
25 )*camxfti arc man kanizak-ra amozum u-cand cust?(
26 ')bh bindum na b(in)daoum be a(z)? nwrbk? yake ::;(
27 b/x)d dahum ta bgydy?-ra be amozao *andarik i siy([ah?]
28 s) kii namah-isuma yaftum be yakc heh az an gufteo '(
29 ') kar i farmiioe-as saxt kunum ti kardah buvao (
30 )hos-um ciz andoh ma xuar ce hos-um-ra xuastah b(w/y
31 )az soy i rwbhh? pursiOum-iOiin? kwpt(y?
32 )pa parvan u-xuao az an soi xuastar-e farmai (sa)r? d(k
33 )andar namah-isuma firestioe kii pa sao u-panj([ah]
34 xua)ndah ?-e an pisez i gospand (a)n? i s(um[a?]
35 8a)hr berftn namaoah-e az soy (i?
36 s ha)mvar? az zen u-rikibain u-duval
37 )az har cc hyz? i barin (az-1
38 -xar?)
31
32
From Old to New Persian
Translation
1 ) the Lord God, who shall be [our] helper, soon the day (
2 ) and more [than] twenty letters I made, but (you'?
3 observe) !that! to what my little bowH attains and to ... (
4 ) order to give his three shares! Also for me (
5 ) ... you buy, until I have set out [and] gone down (
() ) good ... , the Lord God to you for it a good'? reward(
7 ) was delayed until the ninth oH the month and unt.il the tenth (of!
8 )shePp was, and they buy market is] weaker, and the Lord
God ... (
9 )should make that not any of them is worn'?, as they(
10 what'?) clothing that has been sold, this'? towards us/in our face (
11 )had sold: there was'{ nobody, people of the city (
12 t.hose'q thirty jars'{ we should buy, and there is not any nard'! to be
found (
1:3 ) as otherwise I to you, and l have a man, one work(-'?
14 )to know my profit and loss, and 9 *.ubil"i? (
15 likP) sheep to buy on my behalf, so that/until (
rn )thus you said !that!: Rabbi'?, thirty (
17 )to come, and I am/it is heavily attended by loss(
18 ) him from my side 17 bales [of cloth]'?(
19 )that yourself you bought and yourself you sold and yourself (you
strove'?
20 ) if for me profit should be the share, I (
21 ) hut you, for a good reckoning (do) not (suffer) any anxiety(
22 )you sent, and it/he was not here, and the profit of the 1:1hccp thm;'?
correct/ly /(
23 of?) yoms arrives, as God wishes, and you personally near Sb'bd?(
24 )to ( ... ) 8b'bd? ... lthatl: to me one harp (
25 )you bring the harp, 1 shall teach the girl, and how quick/lyf'!(
:W ) .l find, 1 did not find, but from? Nwrbk? one (
27 ) l gin-, so that he shall teach Bgydy?, the eunuch(
28 ) that I received your letter, but you said one better than that
:W ) the work which you ordered, [ shall work hard, so that it shall he
done (
30 )my mind, do not suffer any anxiety that my mind [a.cc.] hurt (
31 )as for the meat? ... J asked thus1 ... (
32 )with licence?, and yourself you are rec1uesting from that party, order
the headl(
The Jewish-Persian Fragment from Dandan-Uiliq
33
33 )in your let.t.er you sent. jmcssagP j that with/t.o on<' hundred and
fifty(
34 )you have (called?) that trifle for the Rheep t.haU of (yours!
35 )city you have not come out, from the side (of?
36 as for saddle and stirrnps and strap:-; (
37 ) from whatever ... of the highest (Az-
38 -xar)
Notes
l. ya,ze)j (H, 8, 23), 'God'; cf. Pahl. yazet (Nnrn1w, A manual of Pahlavi,
1, 1964, p. 150) and NP place names like Yazdi:-c
11
ii,st:
BSOAS 20, p. 342, n. l, reads ized, but. the writing of initial vowel
without ale/ is doubtful in this text (on the possible ydwn ebun,
sec infra 22).
xuaMh/xubah ((), 8), 'Lord'; {i sa,ivtwn (cf. paidiih, 12);
-h, see LAZAHD 1963, p. 172; also, e.g., J:t'. W. K. MOLLElt, Fest-
schrift Sachau, 1915, p. 222b; "Zurn mit.tdpersischen
Passiv", lzv. Imp. Akad. Nauk, 5, rn (HlOO), p. 271.
i, 'who'; i</,afa as rel. pron. is clearly attested in this text (see 10,
29; cf. HENNING, Mitteliranisch, p. 79 f.) as in the Law report from
Ahvaz (ASMUSSEN, Acta Orientalia 29, p. 57); note the writing as a
separate word 'y (cf. HENNING, Mitteliranisch, p. 89); SALJ<:MANN
(p. 049) joins it to the following word to form ayyar: HENNING
(BSOAS 20, p. 341, n. 3) prefers to regard it as an abbreviation
(""N =-= ), seemingly for ilii1" "liN.
2. be/bi, 'but'; here most probably advcrsativc conj. ( - Pahl. BR'//Je,
MMP b') as in 21, 26, 28; SALEMANN (p. 049) suggests a confusion
with pa meaning 'to'; such a prop. by, 'to', whatever its origin may
be, is not uncommon in JP texts (cf. B80AS 31, p.
255), but in the fragmentary material hero discussed there is no
clear example of such a function of by, quite common, on the other
hand, as a verbal particle (4, 7, 8, 14, 27; cf. p. 252);
the writing and the use of pa (3bis, 6, 10, 13, :J2, 3:J; Pahl. PWN/pat,
NP ba/bi; cf. LAZARD 1963, p. 387) are distinctly separated from this.
3. ku (16, 19, 24, 28, 33), 'that' (=Pahl. 'YK/kii; not uncommon in ,JP;
cf. HENNING, Mitteliranisch, p. 79 and ref. in n. 2; ASMUSSEN, AO
29. p. 54; MAcKENZrn, BSOAS 31, p. 252).
jamak?' 'little bowH' (or camak?; cf. also line 10 .,mh c.= jii,m,ah,
'clothing'); a more tempting solution with regard to the meaning is
34
From Old to New Persian
an 1111at.tl'sted jftmnk (with secondary y-/j-), 'po:;t, postman or
post-lwr:;c'!' (NP yfi:rn de., Turk. yarn), giving the approximate
tnt11Hlatio11 'with what my po:;t arrives'; MAROOLIOU'l'II's suggestion
jrl111ayi, stiJ><'rnl' (p. 7fl'2, 7f>5), :;ccms imprn:iHible t.o reconcile with
the writing.
dzt?, name of placP or pcrHon'{ Earlier :;uggcstionH, <last, pu-dast
du.:d, are not satiHfact.ory with regard to the actual writing.
-L 8e/si, 'three', to he distinguislwd from .-;yh-8ih, 'thirty' (12, 16).
'his :;han", probable t.ramilation in the context; 'reaches
him', as sugg<'Hkd hy (p. 050), would be possible if we
could read i11Ht.cad of 8e; not.e the writing of the encl. pron., ys,
as in 1Ja()-i. (6), while 29 seems to cf. LAZARD
19t>3, p. 24fj),
ane ( 13), other, also' ( 0 -cl\1.MP 'ny; cf. al:;o MAcKEXZlE, BSOAS
:11, p. 254).
5. -ybryn?, the fir:;t readable rat.lwr y than w, but the paper is
somewhat folded at the c<lgc.
H. par)-is, 'for it', prep.+ encl. pron. :1. sg. (for the prep. va/pab, sec 2,
note be/bi; cf. also LAZARD um:l, p. 24(), 387).
7. the writing of 'uftab without initial alef is quite
explainable through the connection with be (differently HENNING,
BSOAS 20, p. :342, n. 1).
-nuh i.? ntah, 'the ninth 0'1 the month', suit.H the context well,
but the -irfiifa is not clearly written (more like my!).
8. u-yazeb, the w ii; clearly readable in the original.
:l:yxic?, pPculiar as beginning of a Persian word; perhaps a mistake
for M ;xu(rtluu)), cf. line 2:3.
9. kumW, 3. sg. subj. MP type (sec HENNING, ';\'erbum des Mittelper-
sischen". ZI I 9/1933-34/, p. 2:J3; cf. also RAPP, Die jiidisch-
persisch-hebraischen Inschriften aus Afghanistan, 1965, p. 71; LAZARD
1963 p. 338).
farsilO?, 'worn, decayed'; at least a possible reading-the paper
is here twisted as well as torn and folded.
10. f'ur6xtnh bl.i.r)-a,.'Jt, 'has been :;olcl' 01 'he had sold'; cf. LAZARD 1963,
p. :1.i2.
i.n?. the afof is doubtful: lwn>, aH in many other places, a further
nstoration of tlw paper might he of holp.
l l. bwr. most pmbably a miswriting for bilb/lnwnr).
8aM mardumr711, hun<lrc<l1 people'; the writing 8d 18 probable,
The Jewish-Persian Fragment from Dandan-Uiliq
35
as there is a fold right acrosR t.he sign that. scemH to b<' a single
letter on the photo; fo1 pl. noun after a numeral, sec LAZAIW I 9<.i:l,
p. 218, f.
12. *pitku, 'jar?', from Aram. (I mm this suggestion to H. S.
NYBELm); preferable to 'writing etc.' from Aram. from Greek
mT't'cXX.Lov (also Ruggcstcd by H. 8. NYB.Lm.o) with possible connee-
tions to Turk. bitkii/bitku (cf. VON GABAIN, Altt.iitkische Grammatik2,
1950, p. !104; but cf. YosHJ'J'AKE, BSO[A]H 6/1930-32/, p. 4!) ff.).
e xariJ;ni, 'we should buy'; as already suggested by Salcmann (p.
048), e appems to he the old opta,tivc particle (Pahl. '?/ev, MMP
hyb; according to HALEMANN, Izv. Imp. Akad. Nauk, 5, 13/HlOO/,
p. 275, hy in the ,JP "Commentary on Ezekiel"); xar(f.111, l. pl. snbj.
MP type (HEN"NINO, Vel'bum deH Mittclpel'sischcn, p. 23:1: cf. kuniilJ
supra 9).
na.rd?, 'narcl, nardus'; the first t\\'o lettcrH are badly damaged,
but diHcernihle Ht.rokcs seem to leave nxd and nrd as the only pos-
sibilites (diffol'ent.Iy l\'1AcKENZIE, BHOAH 31, p. 253, cyz); for NP
nard/n.ilrd, sec LAlJlc'lm., Hino-Iranica, rnrn. p. 428, 455: ef. also na.rd,
'backgammon et.c.' and 'trunk', e.g. Burhiin-i qiiti', <'d. Mu'iN,
p. 2127.
nci-paidrJ.11., 'not to be fomul': for the writing, see li1w I


and LAZAHD p. 172.
l:l. ane, see supra 4.
u-rnard, 'and a man': this reading Hcems rather obvious, a]t.hough
the r is somewhat compressecl; unnyd ="u1ni.cl suggested by l\hn.oo-
LIOUTH (p. 753) and SALEMANN (p. 052) is not possible
BSOAS 20, p. :142, n. :1, proposes 1.mnwd for '1"1'11.id).
k<"ir-?, probably a compound with kiir as firHt <'lement, e.g. kii.r-
8<tz (cf. ASMUSSEN, ,Jewish-Persian InH8, p. :1H), kii.r-,izm<f
or kiir-gar.
14. obviously the counted ohjPct. after t , n: a monetary
connected with
15. az 80!J i man-ri"i, 'on my behalf'; for a,z ... rii, cf. LAZALW 196:1,
p. 367 f.; note nian-rii (15, 20) and marii. (4, 24).
H>. rabbi?, seems most probable in spite of the doubts of earlier com-
mentators (e.g. SALEMANN, p. 052, who, however, only UHe<l the
indistinct facsimile in .JRAS 190:1), in that caHe notable as the only
clearly Hebrew element in the fragment.; t.hc original shows rby
as a likely reading and dby as a poRsihlc rNtcling.
36 From Old to New Persian
17. ziyrln.-{lmand, 'att.encle<l by loss'; the remarkable word farf,lij,mand
quoted by LAZAirn 1963, p. 202 ( 166), from a translation of Beriini
indicates the productivity of the suff. -i1mand well into Muslim
times.
18. 1xua, 'him', for vai-nt, (sec LAZAim 1963, p. 224 ff.).
t6p?, 'bale of cloth ?'; the paper being creased at the edge, it is
hard to determine if medial or final p is written; if not final, we may
have a word like t6pal/tubiil, 'filings/bits of metal, esp. copper'
(cf. further Hurlutn-i qiiti', eel. Mu'iN, p. 527), and thus possibly
some type of
19. ranfibe?, 'you strove', or peihaps r<inde, 'you drove'.
:W. 86b, 'prnfit' (subj.) and baxt, 'share' (prcd.).
21.
'))
.........
be, 'hut' see supra 2.
ci,z etc., cf. infra :Jo and MACKENZIE, BSOAS 31, p. 253.
ffri8tUdi, cf. Pahl. fre8tit-/fre8tlit- (NYBERG, Hilfsbuch
1 l, p. 76), MMP prystyd-/pry8t'd- (HENNING, Verbum des Mittel-
persichen, p. 214, 222 f.) and early NP (LAZAHD 1963, p. 265).
ne, cf. LAZAIW H)():J, p. 440 f.
eOar, 'here', cf. l.1AZARD l9fj3, p. 235 f.
860 i, this reading is quite clear, including 'y, where a small
piece of the paper has been turned about 60.
-ibun (24, 31), -Mi1n (cf. line 16)'? i.e. 'thus'; if so, we have here
either the only example in this text of initial vowel written without
ale/ (cf. HENxrxo, BSOAS 20, p. 342, n. l), or some kind of cnclitic
construct.ion.
2:t .<;f/bd? (24), personal name or title? (against this perhaps: nazdik
without -ir.W/a): lVIAHGOLIOu111's reading (p. 754) of this as NP
8ipahbad/-bud and the chronology based on the introduction here
not of tho "lspahbud of Tabaristan" but also of the Arab general
Yazid h. al-Muhallab (p. 747 f.) was rcfutedalreadybySALEMANN (p.
053 f.), and his objections are well supported by the present reading
of this line' x
11
ahab, doubtlessly 'God wishes'); on embassies
from Tabaristan t.o China in 746 and 754 A.D. see, however, CIIA-
VANNEs. Docunwnts sm les 'l'ou-kiue (Tures) occidentaux, 1903,
p. 173 f.
24. --il51111, see supra 22.
kwpyd?, the context seems to require a verb of saying, but this
could hardly he a form of guftan, as suggested by MARGOLIOUTH
(p. 754-) and BAcmm (.JHAS rno:J, p. 7()0: rejected by SALEMANN,
The Jewish-Persian Fragment from Dandan-Uiliq
37
p. 054), considering the regular way in which this verb is written
in H> frfwpty) and 28 (gwptyd); cf. also the obscure kwpty in 31.
*earnxui, 'harp (or similar stringed instrument)'; possibly 8ogdian,
cf. HENNING, Mittcliranisch, p. 80, correcting his "Sogdica, 1048,
p. 36, il8.
25. cu8t?, 'quick! lyJ', seems more probable in the context than a form
beginning ju8t-.
2(). bindwrn, bindahurn, NP forms of MP v-indiidan/vind-; cf.
Mittelpersisch, p. 80; Verbum des Mittelpersischen, p. 199: NYBERG,
.HilfHbuch II, p. 244.
be, 'but'; sec supra 2.
nwrbk?, personal name? (the apparent reading Nur-bak seems
improbable coni-;idcring time and place).
27. bfJ:ljdy?, pen;onai name'? (hardly Turk. biifJ -'r'idi, cf. VON GABA.IN,
Alttiirkischc Grammatik
2
, p. 310).
*andarik, 'eunuch'; cf. 1-IENNINU, Mittelpcrsisch, p. 80: Grm-
SHEVI'l'CH, A grammar of Manichean Sogdian, In54, p. UH.
28. be, 'but', sec impra 2.
29. kli:r i farrnur)e-a./-i., "the work which you ordered'; on i as rcl. prnn.,
see supra line 1, and on the cnclitic -a.9/-i., imc line 4 (ra8nb-i.): cf.
also HENNTNO, Mittelpersisch, p. 80.
30. {;'iz etc., cf. Hupra 21 and MACKENZIE, BSOAS 31, p. 253 (with a
translation).
31. rwbhh?, obscure (hardly connected with r<Jbah, 'fox'); a personal
name'(
kwpty?, looks like kufte, 'you have beaten' (minced t.he nwat.'?),
hut that seems improbable in the context; rather connected with
the obscure kwpyd? (24), if not to be separated kii pty- (against
the writing).
:32. parvii.n, (cf. NP ]Jarvii:nah and SALEMANN, p. 055): Pnl"l'iin
is also the name of a town in Afghanistan (cf. Burhr"in-i qiiW, ed.
M U'IN, p. :392; AHLANOV, Afgansko-russkij slovar', HHW, p. 167 b ).
33. fire8t'i<)e, 'you have sent'; cf. supra 22; 2. pl. with ending sec
LAZA HI> 1963, p. 267 f.
34. 'small piece of money; trifle'; thus another possible translation
is: 'the money for the sheep'.
35. na:ma(5ah-e, '=NP nayiim.adah-i.
36. hanwii.r?, "smooth, equal, suitable, etc.'; -i'i- defective .scn:pturn;
there is a faint possibility of an alternatiw reading: 8ami1r, 'soble'.
38
From Old to New Persian
rikibain, '[pair of] stirrups'; Arab. rikab with imala and dual
oblique ending (final ale/, as read by SALEMANN and MARGOLIOUTH
and referred to, e.g., in LAZARD 1963, p. 157, n. I, is impossible).
37. hyz?, obscure, if not somehow to be read together with ce; see the
interesting chyz in the "Commentary on Joshua" (line B3), inter-
preted by MACKENZIE (BSOAS 31, p. 253) as a writing for NP ciz ...
from MP ce-( i )z.
37 /!38. az-xar?, to judge from the cutting of the paper and the com-
pressed writing of the last two lines, we might here have the end
of the letter; one would expect a signature of some kind-perhaps
the last two (indistinct) lctte1s of 37 and the only two letters (also
indistinct) of 38 together form the name of the writer; a possible
reading is 'zxr, for Azxar (or Azkarp
Addendum
A recent article by E. L. RAPP on "The date of the Judaeo-Persian
inscriptions of Tang-i Azao in central Afghanistan" (East and West
17/1967/, p. 51-58, fig. l; I owe the ref. to R. N. ],RYE) rejects the
ea,rly date assigned by W. B. HENNING to these inscriptions (752/3
A.D.; v. supra p. 125), arguing for the con5iderably much later 1300 A.D.
This may be correct or not. It does not, in may opinion, change the
probable date of the fragment discussed here.
Another recent article of importance in connection with this text is
G. LAZARD's much needed outline of Jewish-Persian dialectology
published in Studies in Bibliography and Booklore, Cincinati, 8 (1968),
p. 77-98 ("La dialectologie du judeo-persan"; v. esp. p. 86).