SILVERJUBILEE VOruME

OF THE

ZINBUN~KAGAKu~KENKYUSYO

. KYOTO UNIVERSITY

KYOTO

1954
 , I
." , f,

"

I ,.
MADU

tl A CONTRIBUTION TO THE HISTORY OF · WINE·

. • by

H. W. BAILEY
" , .
Capili!idge, Universi ty ,,

I
The Iranian madu, originally meaning 'honey','thence used for 'tnead',
later for 'intoxicant drink ',-," wines ,,-
' -' , especially ' wine froni grapes,
nas within re,cent years been found in the language ' of the Sogdian's!, in
manuscripts from Tunhming and Turfan, and in the languages of the
Saka people from Xhotan "and Tillushuq. ' The word (as" it is hoped to
show here) 'was adopted by ' the people of Kucna, who in turIl gave ' it
I

to the Turks. Earlier from' another form 'Of the 'same Iranian word in
Ferghana the Chinese in 128 RC. took the 'name. The present state
of this problem is the · subject .of the following .notes.
I . .

II
Old Iranianl ) *madu is attested by Avestan ma8'(J, 'Wine', and ma8umant..
'containing wine \ epithet or'the Zoroastrian 'myazla- 'offering'. This
word ' is, the well-known Iranian cognate of "Old Iridian madhu 'honey,
honey-drink ~: the~ used of the so~a, the intoxicant of the , sacri'fi'ce. In:
the KautiliY!l Arthasastra2) madhu is explained by . mrdvikiirdsa- 'juice
of grapes'. The Old Indian l1iadhu wasreplaced In KroraiiJ. (Lou-Ian)
north-western Prakrit by the frequent masu · (written so with su, for
*mazu~ since su was not marked with diacritic for zu). Of the related
words of other Indo-European . languages . it will suffice . to note Greek
. . ~ .

,1) "Old Iranian" is here used .for the earliest de!;lucible form ·of .the language attested
in the Avesta and Old Persian itiscriptions of the AchaerneIJians. . .
'. 2) . Ed~ l. ololly and R. Schmidt, p. 71. Cited in my article H(ira"u~ in the Festschrift Fr.
Weller (in the press).
2 H. W. BAILEY

methu 'intoxicant', Kuci mit, mU 'honey', and Germanic, Old Engl.

medu 'mead '.
Iranian *madu is widely attested in later Iranian doyuments. In
Sogdian we find nom. sing. mw8y, acc. sing. m8w, myw, that is, mu8
and ma8u (mayu) (or perhaps mu8u).I) Sogdian gave to Persia mul,
expressed in Armenian as mot/) and from Persian in Romani mol 'wine',
sutw mol 'sloe gin'.
In Ossetic" mud and myd 'honey' has kept the oldest meaning. The
drink made from honey however was known and is called in the tradi­
tional tales rang, which has survived in living use in Mingrel and Svan
rang-. 3)
Khotanese has mau 'intoxi~ant, wine', once written ma in the phrase
ma va-1f£ hara ' bring me wine " translating the Chinese tcyau tta ma le. 4 )
In medical texts mau renders Sansk. madya- 'intoxicant', Tib. chari.
In one textS) occurs ysaril1f£ gilriiT}ai mau 'yellow grape wine '. The
forpl mau has, as, regularly, lost intervocal~c -d-; the -u may repJ;"esent
Old Iran. -u of 1!Ladu.6) The adjectival form is mauya a~d mauva.
T1;le Saka language of Tumshuq1) has {Lhu, almost certainly meaning
'wine '. . If it is rightly so translated, it should probably be coqsidered
as another form of rr;.adu-, but the changes have been greater, and the
stages ·of the change can at present only be conjectured : if ma8u was
replaced by {3a8u ((308u) or y.a8u (IJOOu) , the loss of the initial sound
may be .due to dissimilation from the following _u,S) The ·h- replacing
1) W. B. Henning, BSOAS 10; 98. Sogdian dialects have either h or I from Old Iranian
initial and intervocalic d. New Sogdian in Yaghnab hus d.
2) In mot·ozmay 'wine.glass '. In this ozmay I would see the same base rnli· '~easure,
contain', from which with other preverbs come NPers. paimiinak 'bowl', and Zor. PahI.
patmiin 'bowl' occurring in Greater Bundahisn 62.6; 136.7; Zatspram 6.3, in texts parallel
to yiimak and tast. Khotanese hamau, hamo 'cup, bowl' (in E and Khot. Bud. Texts .p. 7)
will be from the same base *fra.ma.van,. (not with Sten Konow, Primer of Klwtanese Saka, p.
36, from *h{iotna8u,,),
3) V. I. Abaev,. Osetinskij Yazik i Fol'klor I 348 IF. In preference to Abaev's connexi\>n
ofrong with frii,na,. 'bre~th' (known in Sogd. (1r"n), hence' spirit ~ and' spirituous' liquor, I
should seek a cOimexion with Osset.;o; of i.rayun: irad·' 'to seethe, boil' named from fermented
mud', honey'. .
4) Klwtanese Texts H, p. 1, line 5. tcyau is Chin. Mi tsiu • spirit distilled from grain,
wihe from rice '. .
5) Paris Y29.
Q) From, pasu- comes pasa- ' cattle '" Elsewhere we have ·u from ·u-ka· as in ysanil 'knee',
plur. ysiinve from ziinu- wi th ·ka.
7) See . my Languages of the Saka in B. Spuler, Handbuch der Orientalistik; ' Sten Konow;
Ein neuer Saka-Dialekt, and Norsk Tidsskrift for Sproguidenskap 14. 156 IF.
8) As Krorain arnavaj'i may contain varnava-, see BSOAS 11.793. Loss of u;. is regUlar
in Ossetic before i· in the preverb vi·. In uidon, idon, Dig.uidona. idonli 'bridle' both forms .
are recorded. Before u- the {1. hasbelmlost in Oss. ud from bUta-, a8 in Wal]et8i un 'naked'
from Parna-, and in Pasto u¥i •long' from brz-.
 MADU, A CONTRIBUTION TO THE HISTORY OF WINE 3

-8- between vowels recalls the -h- of Khotan. byah-' to change', if :this
is from vi-vad-. '
.Pasto melawd 'grapes' from 'l'miidawya., and Balocl maviC 'raisins'
from *madvi-, have been associated with Jndia,nised Sans](. mrdvika. and
Pali mu4dika•. 1)
In addition to ma8u A vestan has also mada-, ma8a. 'intoxicant' cor·
res.ponding to Ind. mada-, , In Khotanese maya-, instr. meva ,jsa, rende,rs
Sansk. mada- in the SuvarI?-abhasa.!) The related participle is, miista-.
Turfan Mid. Pers. and Parth. my, Zor. Pahl. m'd,:rnY, m'k *ma8.
Krorain me3) are ambiguous in OrigiI:l §!ince , either madu· or mada- wOllld
suffice to explain the form may.

III
The history of the vine has been much investigated.4) For the present
purpose it will suffice to point to the evidence assembled by Schrader
for the knowledge of wine among Indo-European peoples. ' The vine was
known in pre-historic Italy and in Greece in the Mycenaean period. S) It
extended from the Balkans to the Caspian sea, in which region the origin
of the wild vine has been sought. To this evidence can };le added recent
information from Anatolia. , Here three forms of the word for , ' wine'
have been cited~): Hittite in cu,neiform ' script lJi~an-, in hieroglyphic
script lJ:ianas; Luwian lJ:in-. This is the, long-sought oldest cognate of
the Greek oinos, Lat. uinum, Aqnen. gini, Mingrel "rvin-i and Geo~ian
rvm-o.
Early Greek writers knew of the abundant use of wine in Persia. The
A vramiin ,documents 'are concerned with vineyards. 7) ,In the Tale of
Husrau wine is praised and some wines are named: , kangik 'Sogdian',
1) The Bud. Sanskrit Vinaya text hasTTl[dv1kli;panam, see Gilgit 'Manuscripts III l. ji: J.
Charpentier, Acta Orientalia 7.191 dealt with. melaUia, 'The umlautpas9ing oyer a syllable is
like that in Khotan. hiSana' 'iron' from ·as~any~, and ysl~ 'gold' from.arany~. For the
secondary -r- in mrdvikii, note the lavrna- ' salt' for the usual lava~ in the Brihnii inscription
dated in the time' of Huvi~ka, Sten 'Konow, Epigraphia Indica 21 (1931)" p. 60, line 7.
2) Khot. Texts . I 36 ' r 1; P351a, 63 r 4; Khot. Bud. Texts, p_ 63,. 79 r3. , .
3) Frahang iPahlavik 5. l. Krorain me, recognised by Renning; BSOAS 12.603, has-e
which can represent either -e o~ .:ai. , The diphthong ai could be ,shown in Krorain by the
macron stro'kebelow e, as was done .i n sthaira from sthavira, and saila.
4) O. Schrader-A. Nehring, Reallexikon; B. Laufer, Sino-lranica 220 If.; R; Campbell
Thompson, A Dictionary of Assyrian Botany 327 If.
5) Th~ word wo-ne-we :may mean' wine-dealer' in a Mycenaean tablet, see M. Ventris and
J. Chadwick, Journal Hellenic Society 73.97.
6) H. Th. Bossert, ' lahrbuch f.kleinasiat. ,Forschung 2. 180-1. J. Friedrioh, Hethitisches
W'orterbuch 336 has registered ~uia~? and hieroglyphic wG!.i)a~.
7) E. H. Minns, J. Hellenic Society 35.22 fr. '
4 ,H. W BAILEY~ - (

harevik' of' Herat 1, marv-rotik' of Mel'w-rod ',bustik 'of Bust ',hulviinik

'of Hulwan', asorik 'of Asorastan', viizrangik 'of Bazrang'.1)
Wine is attested in 'Afghanistan,: Kashmir, Persia' anp Ferghana. . The
Chi.nese first learnt of wine froID"grapes in -128 B. C. in Ta.yiian;Parthia,
K'ang and Tashkend. In Kucha it was reported in the 4th century A. D.,
and a .sogdian colony of the '7th century named a town from its vines
in the Lob-nor regIon. . The. Chinese however, did not report this wine
among the ·Turks. 2)
The Indo-lranian word for vine and. grapes, driihjii, which is widely
attested in ·modetn Kafirland Indo-Aryan (though not yet pointed out
in Iranian) : Askun dras, Khowar drO(;, Romani driikh, Hindi dakh, but
which is not quoted from the Veda, may perhaps contain a v:rddhi deriva­
tive of a base *drahja-, in which I should propose to see a word con·
nected with Old Engl. deorc, Irish derg 'dark, red', assuming that the
rtame came ftom the dark berries; as occ'urred in the naming of the
'serVic'e berry . . 3)
.­
lo' •• 11
IV '

1 propose now to recognise , a loan-word 'from Iranian in Kuci , mo~

, intoxicant " inst. ,sing. motsa, attested in Pratimok~a and medical texts,
hence the equivalent of-Sanskrit madya-. The adj. derivative is mOfa$$e.4)
The word is then a: Kuci modification, adapted in spelling, of the Iranian
*mori(u)- found; ,for example, in Sogd. mw8.. Two points need to be
clarified if this comparison is correct.
1. Kuci ~ (in form the same as:Brahmi: dh) and trepresenting Iranian
d or' 8. Whatever valu:eteachers" of the Brahmi script in Central Asia
gave to theak~aras, 5)' the writers of .the language of Kuci did not try
• 1) . J. 'M. Unvala; The Pahlavi Text "King Husrauand his Boy.... p. 26 (with . changes).
On : Asorastiin,see E. Honigman" and' A., Maricq, Rech:erches sur les Res gestae divi .Saporis
41-63, . , ,
2) These and other details 'are in Laufer, Sirw-lranica, loco cit.
3) Earlier a connexion has been .sought withwords,nieaning 'berry,.' or 'thorn', see Walde·
Pokorny, Vergl. Wiirterbuch •. 803, 862,and.Pokorny•. Indogerm. ·Etymolog. Wiirterbuch258. The
word driilc~ii has the' same .form , as liik~ii,' lac., in which one ,should see an·s· derivative' of
the ' base rang-' dark colour; red; brown,' connected with Iran. Khotan . .rriisa- 'dark'; Armen.
loan-word erasx-, see JRAS 1953, 95~For .·the 'seiw:ice berry', Lat. sorbum. ~ee Walde-PokQrny;
lOCi cit. II 499. and in my 'article HiirahiiTfG- in the WeBer Volume.
4) A. F. R. Hoernle. Manuscript Remains of Baddhist Literature, p . .258; J. Filliozat.
Fragments ·de texts ·koutcheens . de medecine 'et de . magie; A. J. van Windekens, Lexiqlie ~tymo-
logique des dialectes tokhariens, p. 69. ,
. '5) , It- is notable that for- the language of Tumshuq ,documents· .a new sign was invented
for the voiced occlusive d. _imp~y.ing that the , Briihmi' d :had. another. presumably fricative;
value J.
 MADU, A CONTRIBUTION,tO' THE HISTORY OF WINE 5

to ' distinguish ,two (or more) dental phonemes. Hence 't and ?-' could be
written for foreign t or d (and dh), beside the use "of Brahmi d. Thus
we find caT[uii.i,!", canta,!" 'candana, sandal', wirot 'virodha '; pud.Tiiikte
, Buddha' with 'deva'" putatatte 'Buddhadatta' ',' magat 'Magadha',
samudtar 'samudra '. Similarly occur in the language of Agni magat
, Magadha " pta-nkat 'Buddha', tanaSal 'dartasala:'.
2. -a- in Iran. *mad(u)-; Sogd. mw8~ .. NPers. rnul from , madu-. The
effect of labial p, b, m, w upon an adjacent -a- has , produced a sound
a or u in many words. I) In languages which distinguish '0 from uwe
find o"as in Armen. ' mht ~ 'w>ihe '; rnog, mov; gen.plur. 'mavan 'Magian',
Romani mal' wine '; Greek mau·, Georg~ mog-v-i "Magian '.
The change cun be seen in ' the following cas~s: ' -' -'
L magu- in Old Pers. magu-s, Av. rnayu,2) Greekmagos, Sabsk~
maga-, Syriac mgws. ' The 0 or u occurs ' in Sogd; mwr~ in' mwy-ztw
'massacre of the -Magians,'mwy'neh {;ynh 'Magiun religion '.3) , Uigur
has mwgwc, *moyui(or*muyuc), plur. moyue-lar ',Magi' in the Christian
Iegend. 4) The -c recalls the -c of kiizac 'jar '; which bdongs with Khotan.
kilysa. 'jar, pot' and NPers. kilzah, and of miskic 'cat : ~) In Chinese
the priests of Sulut~l (Zoroaster) are called tf::fc~(K 675, 120) mu-xu,
older miuk.yua,ina document of 1,2 69 A. D.~) ,In the west occur Zor.
PahI. ',mgwk, mgw, ' Inscription of Sahpuhr Sakansah mgw, ZOLPahl.
m{3wpt 'high priest', Pazand mo(3a8, and as loan-word , in Sogdianmy8f3y,
plur. my8{3t' ; NPer.s. muty' Magian '; maubi8. mobed 'chief Magian',
Armen. mag, inogpet, movpet'; Georg. mag-v-i..!)
2. ,Greek margarites 'pearl J. Mid. Parth. mwrg'r'yd,s) NPers. niurvatid.
3. Khotan. bilnaa- 'naked', Osset. bay nag, Wa~etsi-iin frotn*bagna.,
Sogd. f3yn'k, Av. mayna-, Zor. PahI. brahnak. 9)
4. Zor. Pahl., NPers. palang 'panther', Sogd. pwr8nVO)
5. Sogd. mwysk. ,'fly:' t~otp *niaxsika." . Khotan. r!!;ata ,'bee' from
*mWia. 11) , • '

l)$imilar cases are quoted by M. Gram~oni; Traite 'de ihon~tiq~, 'p, 215~:
2) , As also in' vohu, pouiu,. moum, " " '
3) Henning, JRAS 1944, 138. '
4) F. W, K,Miiller, Uigurica [I] 5:-:10,'w. Bang, "Le' M,useon 39 (1926) 43 ft; 'L. OISCllRi,
The Crib of Christ and the Bowl oJ Briddha,JAOS70 (1950) i61-A. ' •
5) Analytiseher Index kiiiiil;; Khotanese in lEAS 1954; 26. For miskit see,I: .Geri;heviteh,
Grammar of Manichean Sogdian., p. 58. " ,, ' '.'
6) Cited in Chavannes and Pelliot, Traite mq,nif:heen, JA 1913. 146. ,
7) The reverse change of -6- before v t6 a is fouild in Abxaz a:.niiigw • shoe' from GeQrg.
mog-v-i • shoe', Zor. Pah!. mok, see JRAS 1954, 34. " " ' " ,
8) Andreas-Henning, Mittelirart. Manidw.ica Ill, p. 58.
9) Asiea, Trans, Philolog. Soc. 1945;7; G. .n4orgenstierne;' nFL '2, 'Index 38*;
10) See BSOAS 11. 782, " " ' h ' ,
11) Gersheviteh, loe. eit" p. 15; BSOAS 10.590. '
6 I • I ":' I H. W. BA.JLEY

6.Sodg. pwrf 'kettle', pwr-t 'cooked '/) Zor. Pahl.; NPers. puxt
, cooked' beside ' *paxua- in Pal§to pox, plur. piixa, Paraci plwk, pok,
Khotan; pq,ha.. '
7. Khotan. b~~-: bula- 'give' from baxs-: baxta·.
8. Khotan. muhu, buhu 'we' from *maxam.
9. Osset. mud,. myd 'honey' from *madu-.
10. Sogd.zwf-' mouth', Chr. Sogd. zwb-, Av. z(Lfan-. 2)
11. , Uigur sumnu beside symnu and s'mnu from Sogd. smnw 'Ahriman,
mara, evil on,e '. S)
12. Uigur suksumur 'cardamom " Sansk. su~mailii, Tib. sug.smel.
13. Uigtir sumur beside sumir, and samir from Sumeru.
14. Osset. burcii, byre 'pepper' from, Turkish, 4) attested in various
dialects,as Balkar ,bure. The source is ultimately Ind. marica-, and the
intermediaries are now known. Khotanese has miri1[tjsya. mirijsya, Sogd.
mr'ynck'" in Arabic script from al-Bairiinl mrj, vocalised marc, . mariC,
Coman burc, Uigur mire and mure, Turkish .in al-Kasrar1 murc Chinese
'*Jlx (K 1303; 54, 1212) muai-lji-tsie, later mei.li.t!fi.
,15. NPeJ;'s. guriiz 'boar', Zor.Pahl. variiz, Av. variiza·..
16. Chor~sri1ian ' wu~ • wife ',. Sogd. w~w ' from vadu, Av. va~u.&)
17. ' . Turk. Kazan maksim, maks'ima • beer from barley without hops',
Osset. .maxsumii, maxsymii • thin beer without hops', Cerkes maxsame,
baxsame, Turkish in al-KaSyaJ;I buxsum 'beer from barley '.') .
18. . Vologeses, place· name Vologesid, besideParthian coin wlgsy,
Zor. Pahl. wlxs, NPers. valiis, baliis, guliis, Armen. vaiars. 7)
For the presence of an Iranian word in Kuci and Agni reference can
be made to amok 'art ';8)
1) Gershevitcli,loc. cit., p. 87.
2) Gerehevi tch; ~oc. ci t., p. 248. .
.3) Tilrkische Turfa';"Te'ite 7.117 and Analytischer Index. .
4) For Turkish, see G. Neineth, Bibliotheca orientaiis hungarica 5.90; for Hungarian b(m,
see Gombocz arid Melich, Magyar etymologiai uotar, p. 491. Later Turkish has Kazan boroe,
Osmanli burn], Kirghjz burn$. Balkar is in Yaziki severnogo Kavkaza I 75. Khotanese miriTflojsya,
mirijsya is frequentoin medical texts. Sogd. mr'YMk' is in Padmacintlima~i 28. For al·Bairiinl,
see Zaki Validi, Kitdb a/"'saidanIJ.h 129.6. B. Laufer, Sino-lranica 374, treats of marica in
Chinese. Uigur is in ; Ral:hmati, lfeikunde der Utguren I ab.d 11. Ai.KliSra~j, Diwiin Ll:!rat
al·turk, ed. Istanbul i 343 (facsipIHe 173.5) has mitre explained by a/,.filfil.
5) Henning, :QSOAS 10.98. .
6) Radloff, Vef&uch, 1999; O$setic Dictionary ;K. Bouda, Zeits. f. vgl. Sprachf. 1938, 180.
on Ossetic and Cerkes. . ' . .
7) H. Hiibschmann, Arm.en. Gram. 79, with other forms. .
8) The west Irani/m dniiik was discussed in Trnns. Philolog. Soc. 1936, 98-101, and an
origin in ham-auk· pointed out. O. Hansen in Tochari!ch·iranische Beziehungen, ZDMG 94 (194.0),
no. 4, by a lapaus cited Lit. m6kyti • to teach', but Baltic Lit. 0 and Lett. a re.present Indo·
Europ. li, and have no connexion with Iran. au (=Indo·Eur. eu, ou or au). The same article
has evidence for other connexions between Iranian and the languages of Ktici and Agni.
 MADU, A CONTRIBU1'ION TO THE HISTORY OF WINE 7

v ,. . r

Probably wit~ . the spread of the knowledge of wine (the most favour­
able ground for a loan-word) the Iranian word madu reached the Turks,
among whom the earlier Chinese reports did not mention wine. . In Uigur
the word' bor wa,s used for 'intoxicant' in ' medical texts ' translated from
Buddhist Sanskrit. The word bor has been, it is hoped' to make probable,
taken ' from Kuci mot. . It is necessary therefore to explain the difference
in sounds. . I

1. bor renders Sansk. madya. 'intoxicant' .1) Related words are borluq
, vineyard " borluq& ar 'gardener',borCi kisi 'drinker '.2)
2. For the initial variation, b varYing with m in a loan·wora, note
buxsum 'beer', maksima, as abOve; bure 'pepper', Uigur mire aild mure,
as above; Armen. loan·word manousak 'violet', Zor. ra,hI. vanafiak,
NPers. banafsah, Arab. manafiaJ, and banafsap) Variation occurs also
in early Indo·Iranian where the Ind. brav- 'speak' has replaced mrav·
retained in Iranian, as three thousand years later, Khowar b'i· 'die'
comes ' from mri·. The words xormzizda, xurmusta have produced kur·
bustan in the language of the Altai Tatars.4) 'There is ' also variation
with original b: thus Uigur has amari 'some, others' from Mid. Iran.,
Pers. a{3iirir according to the Analytischer Index.
3. Final·r in variation with foreign d or 8. 5) With bor from Kuci
I would now compare the bur 'Buddha' of Turkish Uigur burqan, burxan,
both to fortify the derivation of bor and to support ihis interpretation
of the much·discussed bur.')
1) Rachmati, Heilkunde der UihuTefl 11 22, line 28, a translation of the Siddl).as~a;
2) Radlolf, Ver5uch, 1661,1269, 1271; Radlov-Malov, Uigurische Sproclulenkmiilir 373; W.
Bang, Georgpassion 49; P. Pelliot, T,hnng Pao 1914, 453.
3) HubschJnAlln, Annen, Grqm., p. 191. ,
4) This name was used to render Buddhist loom. From Manich. Sogd. xwrmzt came
Uigut qormuzta and Mongol xurmli5ta. The Altai forms are quoted from U. Harva, Folklore
Fellows Communications 52, 141.
5) Varil\tion of rand .z in foreign loan.words is banal in many languages, and the!Je sound8
may also vary with 8 (or the stops.dand t). Some familiar cases were cit~ in BSOAS 11..787,
8S Bud. .5ansk. kiiga~a- . 'paper', in the Chinese-Sanskrit Lexicons kakari; kiikali, Sogd·. k'7~'kh,
Uig. k'gd', NPers. kara8. .others can be seen in Khotan. hallrai 'myrobalan', Zor •.Pahl. halilak.
NPerg. halllah from a Prakrit form of Sansk. harUaki, for which Uigur.shows "ryry *ariri (:with
-y taken from Sogd. -y .from older -'k), and in Kuci arinik; the Khotanese place name Cira,
in Tibetan fila; Uigur $uk5umur' cardamom' with -r from Sansk. $ii~T1/ailii" Khotan. ~u~~l(J,
Tib. $ug-5mel. The r from 8 (also from intervocalic t) is particularly familiar . in Armenian J
10I!n-word~ from Iranian; and in the Iranian dialects of Tati and Kumzari. .
6) A. von SatiH-Holstein, in Radlov, TiSastvustik 141 If. (pur from Chinese); B. L8uf~r,
JAOS 36.390 If. (bur not originally' Buddha ',- largely now antiquated); P. Pelliot, JA 1925.
1. 254 (doubtful); D. Mironow, Rocimik Oriental. 6. 74 (bur from Kuci, by graphic error of
Uigur r for t); Analytischer Index (bur from Chinese); W. Eberhar~" ..BeIleten 35, Turk tarik
kuru mu 1945 (burqan not originally 'Buddha '). . .
 8 I' '. H. W! BAILEY o.

The name' Buddha' was received in the Tarim region before reaching
China. The most immediate source for a Turkish name of 'Buddha'
was therefore the kingdom, ofKuci and Agni. . E~idence f~r the pcipular
form of the name of B}lddha(heside the learned Sanskrit one) is now
~ abundant. Thus we find BOdLJOon the KuSiin coins. In Krorain ' but­
occurs ~n the ' name b,/ ftsena-,butsimga- ' beside the f~ller budhasena~,
budhaseng·a-. Khotan has in compounds prattika~bzida-yauna- 'pratyeka­
buddha-yana.', budii~aittra- 'huddhak~etra~', and in proper names na~
maubudii, saTftgabudi, vidyabudi, siribudi, brabuda, budarma- (for bud­
darma-).I) Kuci has pud in the compound pud-iiCikte (in pr()se pa-) with
the proper name puiatatte 'Buddhadatta', 2) and Agni has pattd-nkat,
ptd-iikiit. Sogdian has Bud.pwty, Man. pwtyy, bwty, abI. bwt'h (=but-),
Zor. Pahl. bwt *but, NPers. but. . Uigur has pwt in the Iormwa . namo
buddMya, and pzpr- in' burqan. Chinese offers ~ (K 47) fa from b'iuCJt,
which was about 800 A. D. wdtt~n' in Tibetan script bur,~bw, ~wur,
and in Khotan~se script hvara~ hvira. 3)
Khotanese used balysa- for' Buddha'.4) With ita title is associated
gyasta- 'worshipped', Av. yazata-, giving the phra,~e gyasta- baIysa"
which is used even in the Triratna formula. This gyasta- translates
Sansk. deva, and like deva is used both of 'gods' and 'kings'. In
Kuci and Agni also occur the compounds pud-nakte and piittd-nkiit where
t.he second Gomponent means 'deva'. .The Turkish bur-qan is best ex­
plained as ~ linguistic, calque based upon Kuci or. Agni , words, with qan
, kIng' to render iiiikte. 5) Turkish Uigur has also the full phrase tiingr~
bur-qanf> with tiingri 'god " which recalls the pleonasm found in Zor.
Pahl. kai Vistasp sdh. . The use ' of tiingriwithqan produced th~ com­
pound tiingrikan. 7) On this view the Turkish bur 'Buddha' did not
come from Chinese f!Jtl fa. Later t}le name burqan 'became imprecise
and could mean 'an idol', as happened to . but in Persia. , The Zoroa­
1) Klwt. Bud. Texts, p. 145, 3 v 2; P 5537. 35; Klwt.Texts II, pp. 20, 26, 28.
' 2) W. Couvreur, Twintigste Vlaams Philologencongres, p:' 91. .
3) ZDMG 91. 35. The , final sound of these Chinese' sylitlbles ending'in a dental (indicated
by -t in Karlgren, alldby -8 in Maspero, BEFEO 20) was in the north-west dialect some r­
sound for which G. Haloun used an inverted r in explaining' for me names in Hedin Khotanese·
Chinese bilinguals, as in namaubuda,written in Chinese~1tN nan-morlr1TW (K 650, 640,637
namrmialL-muat) for Which Haloun gave me nddmrmau-mbo inverte'd r_ '
. -4) Tumshttq Saka has biirs(]", see BSOAS 13.651. The word is a nominal derivative from
braz-, the verbal base whic,h expresses intellectual activity, p6ssibly,as I hope to show elsewhere,
connected with Slavonic blat-.
5) Mironow had already seen this, but had wrOngly cJi:plained, the -r of bur as ·a graphic
mistake fort. '
6) M liller, Uigurica III 54.
7) W. Bang, Ungar. }ahrbuch 5.249.
 MADU, A CONTRIBUTION TO THE HISTORY OF WINE 9

strian author of the Bun-dahisn knew but still in the sense of an Indian
god. I)
To support this comparison of Turkish bur and Kuci pud it should
be noted that other Buddhist technical terms are found in Kuci forms
in Turkish. Two good cases are offered by the following.
1. Kuci anantadii- 'immediate (in reference ·to punishment for evil)"
attested in the plural .anantiirsii-ntu. 2) This is in Uigur "n'nttsyky
*anantarSiki (the -ki represents an Indian form with -aka) and "n'ntrys
*anantaris from *anantadi. 3) The · Bud. Sanskrit word is anantarya".
The Kuci word shows -"5- from Prakrit -rz- replacing _ryy_.4) It contrasts
with the Khotanese modification anantanarya- of the same .Prakrit word.
2. Uigur kz'ry 'the red or yellow ka~aya- robe of the Buddhist
mendicant '. The word with metathesis became Sogd. kr'z'kh, whence
come Uigur kr'~', k'r'~'. Both derive from the equivalent of Krorain
ka~ara-, Kuci ka~ar, Agni ka~ar, kii~iiri.5)
The word burqan cannot be discusSed even thus briefly without con­
sideration also of Uigur bur-sang. Here too bur has earlier been derived
from Chinese ' l9a fa. . But Chinese does not use the compound of this
fa 'Buddha' with fff s()ng to refer to the 'community of mendicants '.6)
Though widelr accepted this derivation from a non-existent Chinese
compound cannot be sustained. 7)
The Sogdian has pwrsnV) 'community of mendicants', with adj.
suffix -anak pwrsnk'n'k. From Sogdian the Turks received their pwrs'nk
and pwrswnk *bursang, bursong. In the Uigur the hendiadys bursang
quwray repeats in quwray 'group' the meaning of sang. The corre­
sponding Bud. Sanskrit word is bhi~u-sangha-, for which two Prakrit
forms are found in Krorain bhic;hu-sarrtga- and bhighu-sarrtgha-. The
form with -c;h- (retroflex ch) is north-western Prakrit, and the one with
-gh- from Central IndianPrakrit bhikkhu-. From this same Prakrit come

1) BSOS 6. 279 ff., for him a d~mon.

2) So to b~ divided, not with E. Sieg, Tocharischk Sprachreste, B, p. 89 as anaman-.
3) The. final syllable of Uigur pwdystb was similarly affected by metathesis to pwdysbt
.bodisatfj 'bodhisattva'. References are Muller, Uigurica III 52, 53; Tilrkische Turfan-Texte
4.7; Analytischer Index s. v. anantris.
4) See Trans. Philolog. Soc. 194·7. 142 ff. )
5) It is not ill this word an or· due to the presence of a retroflex sound, as in Gershevitch,
Grammar of Manich. Sogdian, p. 54. For k~ara- see BSOAS 13.130. The Turkish word occurs
in Muller, Uigurica, III 53, 55, 57 and A. von Gabain, Briefe der uigurischen Hilen-Tsang-Bio­
graphie, p. 30. Sogdian kr'z'kh is in the Visvantara Jiitaka, 1497.
6) This, which I had observed, is · confirmed to me by a letter of P. Derrlieville.
7) BSOS 6.280, after F. A. Rosenberg, Bull. Akad. Nauk SSSR 1927, 1394; Analytischer
Index; W. Bang, Ungar. Jahrbuch 5.249.
8) References collected by Rosenberg, loc.cit.
10 H. W. BAILEY I

also Khotan.. bilsangga-, later bisa1'!1-ga- (with adj. bisa1'!1-giTJ,ai, and in

hendiadys, with g~i!:.- 'group', in bisa1'!1-Bije g~), and Agni pis-sank.!)
. The Sogdian pwr-, the Khotan. bil- and the Agni pis- occur where
the. Indian original had bhik~u or derived form. Hence most likely these
three forms are the same word. We should assume an intermediate
*bis-sanga- or *bik-sanga-. This can be supported by the two. following
considerations : ­
1. bu- replacing bi- is a commonplace.. Beside Uig~r mire (from
Indian marica-, as above) we havemurc and burl;. Both bUmll] and
butmll,l occur in Uigur medical texts, derived from Indian pippala-mula­
'root of pepper' by way of ·i(pit(pat),mul. 2 ) Similarly Orkhon Turkish
hasbuqaraq , Bukhara ' (if from vihara-), and Mongol hasbuqar 'vihara-,
monastery' (for which Uigur uses the Sogdian brx'r).
2. the .first of Prakrit doubled consonants is dissimilated .by l, r, d,
t. Thus once3) in the Bud. Sanskrit of Khotan pitpali is used. Such
an Indian form. lies at the base of Uigur ptpdy 'pepper'. The ,Pers.
-Arab. pilpil, filfil, tul/ul 'pepper' indicates the same replacement of -pp-.
Direct from an Indian pippali comes the Turkish bibli in al-KasrarI.4)
Chinese also has -tp- (-8p-) in • ;t,t,. tapi-po from piet-b'uat,piet-puat.5)
Sansk. guggulu-, gulgulu-'bdellium', Tib. gu-gul, is in Khotan Burgula­
and in Kuci kurkal.
The-ls- of Khotan. bilsangga- may represent this same replacement
of aPrakrit oSSa, but it may be noted that in Iranian words I;I.lso Khotanese
has replaced sand z by l in n,altsuta-' gone. out' and naljsema- 'to
finish' which come from nis-eyuta- and niz-Jamaya-.

, . VI
The Chinese learnt to know grapes and wine in Ferghanain 128 B. c.S)
This was a region of the Saka tribes, para SUBdam of the Achaemenian
inscriptions.') The Chines~ adopted the local name for the newly dis­
covered vine and its fruit. This word they wrote fFIf ;fJk, frWla, lrIi·~ (K.,
Gram. serica 102, 1145, 1047) p'u-t'au ftomb'u-d'au. It has long been
1) For Kuci I know only siihk and siin. .

2) Rachmati, Heilkunde der Uiguren n.

3) Khot. Texts I 166, 82 v 5 beside the usual pipali· and Ind. pippali. .

4) Ed. Istanbul, index, p. 90.

5) B. Laufer, Smo-Ironica 375. The Tibetan pi-spal is modified according to Tibetan

habits. .

. 6) Laufer, Sino-Iranica, 221.

7) E. Herzfeld, Archaeologische Mitteilungen aus Iran 4. 10.

 MADU, A CONTRIBUTION TO THE HISTRY OF WINE 11

proposed to trace in the name an Iranian word. I) But only recent dis­
coveries of Iranian materials have shown how the borrowed word can be
interpreted. The pronunciation of the signs as .*b'wo-d'og about 700
B. c., and as *b'uo-d'au .in 600 A. D. allow .the assumption of an inter­
mediate *bodau or *budau (or with fricative *{3odau, *(3udau). This will
give another example of the stop b (or fricative (3) replacing the nasal
m, as known in Ossetic burcii and as claimed above in Uigur bor. The
second syllable is to be compared with the -au- attested from u-stems
in the Avestan aranaum 'fight', Old Pers. dahyauS 'district', and in
Sogd. (Chr.) dyx'w *dixaw, that is, a form arising by ablaut in the
inflexion. It should however also be pointed' that from -u-stems . ad­
jectival .derivatives were made in -a1f:a- and -ayp,-, as found in Khotahese
rraysau- 'empty.' from the base raz- 'be void', and in various Sogdian
sufJixes. 2) . The f<:>rm of *bodau thus understood confirms the claim for
. Iranian origin.
The subsequent development of viticulture in China is sketched
La~fer's book . . From China the name came as budau to Japan.

I '

, I

. ..)

1) Laufer, Sino,[ranica 225; J. Charpentier, Acta Orient. 7. 191; Menning, :BSOAS 10. 98.
No prima hcie case can be m'lde out for Greek b6trus; it l'I'as rejected ' by Laufer, loco cit.
2) Gershevitch, Grammar, p. 164 If.
~:,