The Mount of the Immortals A note on Tamil cultural inßuence in fifth-century Indodrlna

by Paul Wheatley (Chicago)

Ever since Pelliot first drew attention to it in 1903 the Mount of the Immortals has stubbornly resisted identification of both its name and its location 1• The circumstances in whidl it is mentioned are as follows. In the 2nd year of the Yung-ming period of the Southern Ch'i dynasty (A. D. 484), King Kau~Q.inya J ayavarman of the kingdom that is today rather in­accurately termed Fu-nan or • B'iu-ngm 1112 sentthe IndianBuddhistNägasena as his emissary to seek the support of the Chinese emperor forawar against Campä. While in China the envoy gave an account of conditions in • B'iu­n~, a substantial part of whi<h has been preserved in the Nan-Ch'i Shu, 58, 1 Ob-11 a. The relevant passage in this account runs:

• Nä-ka-siän 161 [Nägasena] stated that it was the custom of that country to worship the celestial god • Mua-xiei-si~u-ld 171 [Mahesvara]. This deity regularly descends on Mount • Mud-tgm 181 [so that] the climate is constantly mild and herbs and trees do not wither.

1 Paul PELLIOT, "Le Fou-nan", Bulletin de J'Eco/e Francaise d'Extreme-Orient, tome 3 (1903), pp. 248-303. Earlier in the same year Etienne AYMONIER had attempted to identify the kingdom of Fu-nan but had not mentioned its sacred mountain ["'Le Founan", Journal Asiatique, 10e serie, tome 1 (1903), pp. 109-150). This paperalso contained a Iist of previous authors who had speculated on the location of Fu-nan (including WILFORD, BowRING, GARNIER, REMUSAT, DE GuiGNES, DE RosNY, KLAPROTH, PAUTIER, HERVEY DE SAINT-DENIS, BARTH, ßLAGDEN, SCHLEGEL, and TAKAKUSU) but, as far as I have been able to ascertain, none of these authors had commented on the mons sacer.

2 Fu-nan is the Modern Standard Chinese vocalization of Ancient Chinese • B'iu­n~m ( Old Khmer [Vral;} Bna.I!l = the (Sacred] Mountain, for which see Etienne AYMONIER, Le Cambodge, tome 1 (Paris 1900), p. 283. The dynastic style of the rulers of a kingdom in the lower Mekong valley included the phrase King (or Great King, possibly ( Skt. Mahäräja) of the [Sacred] Mountain: Khmer kuruil bnCllll (cp. Skt. Parvatabhüpäla or Sailaräja). It is usually asserted that Chinese envoys mistook the last term in the Khmer form of this royal title for the name of the country [Louis FrNoT, "Sur quelques traditions indochinoises", Mt?langes d'Indianisme offerts par ses eH:ves d M. Sylvain Levi (Paris, 1911), p. 203, and "Seance du 14 janvier 1927", Jo'!-rnal Asiatique, tome 210 (1927), p. 186; George CoEo:Es, "On the origin of the SaUendras of Indonesia", Journal of the Greater India Society, vol. 1 (1934), p. 67, and Les etats hindouises d'Indochine et d'lndonesie (Paris, 1964), p. 74], though there seems to be no compelling reason why bnCU11 could not also have constitued an element in the name of the kingdom. • 'fhe ~ual Chinese transcription of the name used to denote this kingdom was B lU·n~m but a possible • Piu-ndm (1) occurs once, in Tso Szü's San-tu Fu 121 (3rd

century), and I CmNo, an able philologist whose protracted Iabors in the precise tr~s.lation of difficult mantras bad endowed him with a rare expertise in the trans­cr!phon of foreign words, on two occasions used respectively the forms • B'udt­n~m I' I and • Pud-ngm 141, while retaining the customary transcription in a third reference: Nan-hai Chi-kuei Nei-fa Chuan, I, 4a, and Ta-T'ang Hsi-yü Ch'iu-fa Kao­seng Chuan rsJ, I, 2.

c 1 J ~-m c 2 J -=.t~JIÄ c 3 J mm c 4 J ~m c s J ft?J: mm~JWViii1t, *~g§~3J<~jWj1tJf.t ( 6 ) JJß{hnfw (7) ~uttm csJ 11ttt

91

Nägasena also submitted a written report on • B'iu-n9m, part of which reads:

Good tortune is pervading the world, profoundly affecting all man­kind, the reason being that divine power is bringing about the transfor­mation to enlightenment 3• On the mount of the immortals [hsien shan] 4

called • Mud-t(lm the tree of tortune bears an abundance of fine blossom [(or) auspicious trees flourish in great abundance (chi-shu fu-chia­yung 1121)], so that the divine • Mud-xiei-si~u-ld regards it with favor [as a place on which] to send down his holy spirit. The princes of the country all receive its protection and the people are all tranquil. It is because this grace is all-pervading that the subjects [of the King] are submissive to authority 5•

As long ago as 1928 George Coedes identified the hsien shan mentioned in these texts with the Ba-Phnoxp., a low hill situated a few kilometres to the southeast of present-day PhnoiP-PenQ., but made no attempt to reconstruct the original toponym, presumably an honorific of some sort, that the Chinese rendered as • Mud-t(lm 8• Some thirty years later Louis MaUeret suggested

a These lines are inverse, and PELLlOT "a vu malla suitedes idees" ("LeFou-nan", p. 260, note 4). It seems that the language carries some of the tedmical implications of Buddhist terminology, which is not altogether surprising in light of the fact that Nägasena was a Buddhist and that his report was addressed to an emperor weil disposed towards that faith. Buddhist usage is apparent in the terms hua yüan 181, defined by W. E. SooTHILL and L. Hooous as, "The cause of a Buddha's or bodhi­sattva's coming to the world, i. e. the transformation of the living" [A dictionary ol Chinese Buddhist terms (Revised edition, T'ai-pei, N. D.), p. 142], and ming 1101, presumably to signify Buddhist enlightenment (SooTHILL and Hooous, Dictionary, p. 262), while chi (sri), li (patu, tik~na), kan, she, chün, and sheng [U] are all characters that would have flowed easily from the brush of a devout 5th-century Buddhist monk.

4 Hsien is a word with broad connotations that are capable of a rather wide range of translations. In English it has often been rendered as "(Taoist] immortal". Edward H. SCHAFER normally translates it as "transcendent", and sometimes as "sylph" · He writes, "The word actually connotes 'having the ability to spring up out of the mire of the material world into the realm of spirit, symbolized by unearthly journeys to sky castles and island paradises', with linguistic cognates meaning 'soari~g, flying, flapping up'. In Han times it had ordinarily been used as a verb to charactenze the activity of the 'Feathered Men', who were Taoist adepts and initiates, trans­formed men shown pictorially with angelic wings" [The VermilionBird. T'ang images of the south (Berkeley and Los Angeles, 1967), p. 7]. In the present text I think hsien probably signifies little more than "supernatural• in a vague sense.

5 The concluding pbrase kuei-ch'ing [11) is uncommon, and I have found it recorded in only one dictionary, MoROHASHI 6439/ i where it is rendered as "to wisb toreturn (e. g., home) •. However, the phrase is also cited as synonymaus witb kuei hsin [11), which Tz'ü Hai and Tz'ü Yüan botb explain as "to submit witb pleasure•. It would seem that it was a meaning close to this that was intended in the present instance.

• George CoEDES, •Etudes cambodgiennes: la tradition genealogique des premiers rois d 'Ailkor•, Bulletin de J'Ecole Francalse d'Extreme-Orient, tome 28 (1928), P: 128. Colonel G. E. GERINI came close to the same conclusion wben he equated the Chmese

( 9) 1t . ( 13) lWffl

98

c1oJ aJJ c11J w, f!J, ~. •· lf, ~ c12J -awtta~ ( 14) J;,c,,

that the triple-peaked Nui Ba-the 7, at the foot of which the city now known as Oc-eo was laid out, might have been the sacred mountain referred to, but he similarly refused to speculate on the form of the ancient name 8• At mudl the same time Mme Eveline Poree-Maspero proposed that it should be equated with the hill of Tay-ninh in the Trans-Bassac 9• In recent times this eminence has been known variously, to the Vi~tnamese as Nui Ba-den (= Mountain of the Black Lady), to the Cambodians as Phnm:p Mi (or Me) Den or PhnoiJl Con Ba Den, and perhaps also as Con Bäk Den 10• According to a local legend Mount Me Den 11 (which is the form on which Mme Poree­Maspero has based her identification) was erected by a certain Me Dang 12

during a contest between herseit and a youth from a neighboring village. Versionsofthis myth are widely diffused in Southeast Asia, and are some­times taken to preserve archetyped remembrances of a postulated, but so far undocumented, conversion from a matriarchal to a patriarchal family system 13• Mount Me Den has not only proved to be the site of several ancient liriga, but is - or at least was when political conditions permitted - also a place of pilgrimage for barren women desiring children. "11 faudrait", comments Mme Poree-Maspero, "supposer que Mahec;vara [of the Chinese text) 'descend' et a son symbole sur le mont bäti par les femmes, et qu'inversement Umä [Siva's consort] serait honore sur le mont dresse par les hommes," that is the Ba Phnm:p of which Umä Mahi~äsura is, in Cambo­dian folk Iore, the guardian deity (niik ta), and whidl was popularly alleged to have been raised by men u.

Essentially this argument rests on an implied phonological similarity be­tween the Modern Cambodian Me Den and the Ancient Chinese • Mua-t~m (though Mme Poree-Maspero cited the graphs in their Modern Standard Chinese transcription as Mo-tan). At first glance there is an attractive simpli­city to this reasoning which promises simultaneously to provide answers to

graphs for Fu-nan with the Khmer phnorp = hill, and located the capital of the country at Ba-PhnOip: Researches on Ptolemy's geography of eastern Asia (London, 1909), p. 209.

P_7 Known to the Cambodians as Bhnaril Pä The, Pä Dhe, or sometimes as Bhnaril ad Salimer.

8• Louis MALLEBET, L'archeologie du delta du Mekong: I, L'exploration archeo­

l(Pog1f!Ue et I es fouilles d'Oc-eo (Paris, 1959), p. 75, and 111, La culture du Fou-nan ans, 1962), p. 453.

(p• .Eveline PoBfE-MASPERO, Etude sur Jes rites agraires des Cambodgiens, tome ~;1s, 1962), pp. 139-140.

1 Cf. G JANNEAu, Manuel pratique de Ja Iangue cambodgienne (Saigon, 1870),

p. 42. 1.1

Me is a common Cambodian appellation for both women and the females of anliD.als [ cf. Franc;ois MARTINI, •oe Ia signification de 'Ba' et 'Me' affixes aux noms de rnonuments khmers•, Bulletin de J'Ecole Francalse d'Extreme-Orient, tome 44 (1947 ~),PP· 201-209), but den carries no recognizable connotation.

These forms illustrate a practice of maintaining alliterative affinities between ;ames connected with the same event that is common in southem Indochina: v ide

OME-MAsPERo, Etude, p. 139. u Louis MALLEBET, •Traditions Iegendaires des Cambodgiens de Cochinchine

~el~;rant d'une interpretation ethno-sociologique•, Institut lndodlinois pou.r l'Etude eu Homm.e, Bulletinset Travaux, tome 4, fascicules 1 et 2 (1942), pp. 169-180.

PoriE-MAsrEB.o, Etude sur Ies rites agraires, tome 1, p. 139.

99

the twin questions of the identification and location of the sacred mountain But on closer inspection there can be no doubt that it is a misleading over­simplification. In the first place, as a tool of the transcriber • mud was nor­mally used to render ma- or -ma- in words such as Mahesvara (a dassie example occurs in the same passage in which Mud-t9In is mentioned); Mahäsri r Mud-xd-siat-lji 1151) in numerous Buddhist texts; Malayu ("Mud­ld-i(}uj 1161 in Ts'e-fu Yüan-kuei, 970, lOa, and T'ang Hui-yao, 100, 13b; Mahä­campä r Mud-xd-tsiäm-pud) 1171 in Hsüan Tsang's Ta-T'ang Hsi-yü Chi, 10, 51a; or Vima (" lwi-mud, for Vima[lakirti]) 1181 in I Ching's Ta-T'ang Hsi-yü

Ch'iu-fa Kao-seng Chuan, 98b. Another transcription in this lasttext(57a) has particular relevance to the present enquiry, for it incorporates both the cha­racters with which we are here concerned, though in reverse order: Tämalitti (the Päli form of Tämralipti) is rendered as • Tgm-mud-lj(}p-tiei 1191. It does not appear likely that a knowledgeable and scholarly Chinese author of the 6th century would have employed • mud as a transcription of the Old Khmer equivalents of Me or Mi, for which he would almost certainly have preferred a form such as • mji~ or • mjet 1201 (e. g. • kiu-liu-miet-ka !211 for karmika: Sui

Shu, 82, 4a). The second element in the Chinese transcription under consideration

does not fi.gure among Karlgren's reconstructions, but presumably belonged with a series of four linguistic cognates !221 vocalized with a sound very close to tgm 15, among them two graphs !241 which have sometimes, though erroneously, been regarded as interchangeable with the present character. None of these graphs, I think, would have been pressed into service by a competent transcriber of the 6th century in order to convey the Khmer sound den. In short, and discounting the fact that Mme Poree-Maspero"s equation is based on contemporary vocalizations, the phonetic correspondance be­tween the Khmer and Chinese names falls short of that customarily attained in medieval Chinese transcriptions of foreign toponyms.

In attempting to elicit a phonetically more satisfactory identification of • Mud-tgrn, it may be helpful to state briefly some of the implications of the texts quoted above in so far as they can be interred on general grounds. The fundamental import of Nägasena's memorial so far as it concerns this enquiry would seem to be that a Saivite cult was practised on a sacred "mountain" - in reality no elevation in the lower Mekong valley could

15 A filth morpheme was slightly irregular, being vocalized as tdm !231 {IÜRLGBEN, Grammata Serica Recensa 656k).

(15) ~~iiJ~ flj

( 19) Jtt'*JL:lif (23) *t

100

(16) lftll il!i­(20) ., ~

( 24) llt ' 1ft

( 17) *~pt:1t

(21) ifij-~~

(18) ~

(22) tt, nt, Jft, ft

have been anything more than a hill - in • B'iu-n9Jn. However, this at first sight seemingly incontrovertible reading of the text has been disputed. Professor Kalyan Kumar Sarkar has argued that the Mahesvara mentioned in the Nan-Ch'i Shu referred not to Siva but to the Bodhisattva specified later in the same memorial, and that the whole passage was a Mahäyäna Buddhist formulation 16• This interpretation, whid:l requires that Nägasena's account be read from a Tantric point of view, is rendered dubious by the date of Nägasena's journey to China, namely 484, whereas the earliest dated evidence of Mahäyäna Buddhism anywhere in Southeast Asia is an inscrip­tion from Talang Tuwo, near Palembang in Sumatra, that has been ascribed to 684 17• The doctrine seems to have diffused at all widely only during the second half of the 8th century in response to the dual stimuli of the political prestige of the Päla dynasty and the teadl.ing of the university of Nälanda. In Cambodia the first extant epigraphical record of Mahäyäna Buddhism is an inscription found in the province of Siemreap and dated no earlier than 791 18• In any case the two sections of the memorial that refer respectively to Mahesvara and the Bodhisattva are cast in different conceptual moulds, and it would seem a priori unlikely that there would have been an alterna­tion of terms between them such as is envisaged by Professor Sarkar. In the second half of the passage the Bodhisattva is referred to explicitly enough as • B'uo-$iil [for • B'uo-d'iei-$Cil-tud l25l] and I have been unable to detect Tantric overtones in the language of the memorial. Furthermore, Professor BhattadJ.arya has pointed out that, in the official response to Nägasena's petition, the Ch'i Emperor admowledged the role of Mahesvara as a source of spiritual power (ling l26l) and expressed approval of the Saivite cult prac­tices, even though they were "alien usages and strange customs" (shu-su i-hua 1271). "Que pouvaient etre," asks Professor Bhattad:larya, "aux yeux des Chinois, ces 'coutumes lointaines', ces 'moeurs etrangeres', sinon des

18 Kalyan Kumar SARKAR, "Mahäyäna Buddhism in Fu-nan", Sino-Indian Studies, vol. 5, no. 1 (1955), pp. 69-75.

17 G~orge CoEots, "Les inscriptions malaises de <;rivijaya", Bulletin de 1' Ecole Franca1se d'Extreme-Orient, tome 30 (1930}, pp. 29-80; Gabriel F ERRAND, "Quatre textes epigraphiques malayo-sanskrits de Sumatra et de Banka .. I Journal Asiatique, tome 221 (1932}, pp. 271-326; Louis de la VALLEE-PoussiN, ltCorrespondance", Bulletin de. l'~cole Francaise d'Extreme-Orient, tome 30 (1930), p. 656. The precise date of ~hts ~ns~ription was established by Louis-Charles DAMAlS, "Liste des principales mscnphons datees de l'lndonesie", Bulletin de J'Ecole Fram;aise d'Extreme-Orient, ~ome 46 (1952), p. 98. 1t is true that this first extant reference to Mahäyäna Buddhism ~n S?u~east Asia does seem to imply a degree of familiarity with Tantric thought, .ut lt 1s a full two centuries later than Nägasena's embassy: CoEots, "Les inscrip­

t(~~s ma~a.ises", pp. 5~57, and Les etats hindouises d'Indochine et d'Indonesie ~ud e~tlon, Paris, 1964}, p. 160,-note 3. .

.Lou1~ FINoT, "Lokec;vara en Indodtine", in G. VAN OEST (ed.), Etudes As1atiques bu~llees a 1' occasion du vingt-cinquieme anniversaire de 1' Ecole Francaise d' Extreme­

nent, tome 1 (Paris, 1925), p. 235.

(26) II: (27) ~1ttf4ft

101

coutumes, des moeurs non bouddhiques" 11? But the primary reason for rejecting Professor Sarkar's thesiswill become apparent later in this paper.

The existence of a Saivite cult in • B'iu-ngm at the end of the 5th century, by contrast, not only appears to be a reasonable inference if it is seen as a precursor of the Saivite devotionalism that flowered in Cambodia in the 7th century 20, but is supported by a passage in the Liang Shu, 54, Sb that pro­bably relates to the reign of Kalll)<;iinya Jayavarman at the turn of the 5th and 6th centuries, but which cannot in any case be later than the middle of the latter century. This passage notes that among the statuary of • B'iu-n{)m were bronze figures with two faces and four arms that can only have been representations of Harihara, a syncretism of Siva and Vi~QU tt. No sculpture of this type from the 6th century has yet come to light, but the stone head of such a statue dated to about 650 has been found on the Phno:Ql Ba-the n. Already by the second half of the fourth century, in the neighboring kingdom of Campä, inscriptions cut to the order of a King Bhadravarman attest to the dominance at the royal court of the cult of Siva-Umä 23, and in the pre­Ankorian kingdom known to the Chinese as • Tsien-Idp (MSC = Chen-la 128l) Saivism was sufficiently prominent to be properly regarded as the state religion. Although even before the end of the 6th century the brothers Bhavavarman and Citrasena-Mahendravarman appear to have been, in Bhattacharya's phrase, "de fervents adorateurs de Siva" 24, the earliest extant

1t Kamaleswar BHATIACHARYA, Les religions brahmaniques dans l'ancien Cam­bodge d'apres J'epigraphie et l'iconographie. Publication de l'Ecole Fran~aise d'Ex­treme-Orient XLIX (Paris, 1961), p. 13.

20 BHATIACHARYA, Les religions brahmaniques, pp. 43--95, and •Religious specula­Uon in ancient Cambodia" , in Himansu Bhusan SARKAR (ed.), R. C. Majumdar Felicita­tion Volume (Calcutta, 1970), pp. 78-97; 0. W. WoLTERS, "Khmer ·mnduism' in the seventh century". Paper read at the London Colloquy on Early Southeast Asia, The School of Oriental and African Studies, University of London, September 1973.

11 Pierre DUPONT, La statuaire preangkorienne. Artibus Asiae Supplement XV (Ascona, 1955), p. 21 j CoEDES, Les etats hindouises, p. 120.

11 Louis MALLERET, Pour comprendre Ja sculpture bouddhique et brahmanique en Indochine (Saigon, 1942), pl. XXVI, 1, and L'archeologie du delta du Mekong, I, PP· 89 and 409-410, and pl. LXXXVIIb; DUPONT, La statuaire preangkorienne, p. 28, pl. 111 A.

13 Auguste BARTH and Abel BERGAIGNE, Inscriptions sanscrites de Campä et du Cambodge. Notices et Extraits des Manuscrits de la Bibliotheque Nationale XXVII, 1 (Paris, 1885-93), p. 199; Louis FINOT, •Notes d'epigraphie, I: Deux nouvelles ins· criptions de Bhadravarman Ier, roi de Champa: Bulletin de I'Ecole Franc;aise d'E_x­treme-Orient, tome 2 (1902}, pp. 185-187, and •L'inscription de Chiem-son•, Joc. c1t., tome 18 (1918), pp. 13--14; R. C. MAJUMDAR, Ancient Indian Colonies in the Far East, 1: Champa (Lahore, 1927), Inscrip. No. 2; B. C. CHHABRA, •Expansion of Indo-Aryan culture during Pallava rule•, Journal of the Asiatic Society of Bengal: Letters, vol. _1 (1935), pp. 47 et seq.; George CoEn:Es, •La plus ancienne inscription en Iangue chame • New Indian Antiquary, extra ser. I (1939), pp. 46-49.

u BHATTACHARYA, Les religions brahmaniques, p. 24. In the Sui Shu, 82,_ 7b-8a, whic.b. is here reporting on events immediately prior to 589, there is menbon of a mountain called the Liitgaparvata (Chinese transcription = • Liang-ka;puan­b'ud) l"l, on the summit of which was a temple dedicated to the god (shen) • Bua-ta-1}1 l3 I == &hadre[svara], the name under which Siva was worshipped at Vät Phu (COEDE , •ta tradition genealogique des premiers rois d'Angkor•, p. 124]. It was during this period, too, that Citrasena, later King Mahendravarman, erected a

(28) (29) ~{bß~~

102

explicit expression of Saivite doctrine in Cambodia is incorporated in an epigraphic invocation to the god recovered from Phnoq1 Bayail and dated 526 Saka (A.D. 604) 25• Nearly a quarter of a century later (549 SI A.D. 627) the earliest mention of the Päsupata sect occurs in an inscription from Sam­bor-Frei Kii.k 211 • To sum up: even though there is no extant epigraphic evidence of Saivism in • B'iu-n9m contemporary with Nägasena's petitionl sudl cults became so prominent in -succeeding centuries that it would seem not unreasonable to regard the Mahesvara of Mount Mua-t9ID as an early manifestation of the same complex of beliefs.

The Chinesetext also exhibits an interesting parallel with a ninth-century Javanese metrical inscription transcribed1 translated1 and annotated by de Casparis 27• In the translation of the textus amplior on p. 11 I have proposed two possible translations of chi-shu lu-chia yung. Pelliot offered a version similar to the second of my alternatives ("les arbres fortunes y prosperent en grand nombre") 28• Howeverl if the reference were to a single tree, then the passage would be reminiscent of the following lines abstracted from a rather lengthy description (perhaps evocation were a more apt word) of a Javanese Saivite temple-complex:

" ... there also was a Taftjung tree ... Of the tree Ki Muhür (?) the trunk was only one year old. The proximity of the Lord was the reason for its matdJ.less growth on the Eastern side. Its beauty was extraordinaryl equal to the [divine] Pärijätaka tree; it was the place where the God would descend 29 and [its brandJ.es] would be a parasol (for the God]. Was it not a God for the God 80?

The Pärijäta (or Pärijätaka) to which the temple tafijung tree (Mimusops elengi, Linn.) is here compared was one of the five sacred trees produced

series of Jiriga along the Mekong, in the vicinities of Kratie and Stung Treng (Louis ~rNoT, "Notes d'epigraphie1 IV: Inscription de Thma Kre (Cambodge) "I Bulletin de ! Ecole Francaise d'Extreme Orient, tome 3 (1903), p. 212; Adhemard LECLERE,

Une campagne archeologique au Cambodge" 1 loc. cit., tome 4 (1904)1 p. 739], and in the region between the Mun river and the Dailrek upland [Erik SEIDEN­FADEN, "Complement a L'Inventaire descriptil des monuments du Cambodge pour ~es qua.tre provinces du Siam oriental"1 Joc. cit., tome 22 (1922), p. 57-581 and Chromque", Joc. cit., p . 385].

25 BARTH and BERGAIGNE, Inscriptions sanscrites, V, cf. also Kamaleswar BHATTA­CHARYA," La secte des Pä~upata dans l'ancien Cambodge", Journal Aslatique, tome 243 (1955), pp. 479-490. ~ George CoEnts, In.scriptions du Cambodge (Hanoi1 1937) tome 1, pp. 3 et seq.

7• J . G. DE CASPARIS, Prasasti Indonesla, vol. 2 (Bandung, 1956)1 pp. 280-330. For

earher comments on this inscriptions see J . L. A . BRANDES, Catalogus Groeneveldt (l~!). P· 382; Nieholaas J. KRoM, Hindoe-Javaansche Gesdtledenis (Second revised

. e?-ition, 's-Gravenhage), p. 138; Louis-Charles DAMAlS, "Etudes d'epigraphie indone­~.1enne , 111: Liste des principales inscriptions datees de l'Indonesie" I Bulletin de E~ole Francaise d'Extreme-Orient, tome 46, fase. 1 (1952), pp. 1-105. n PE~LIOT, •Le Fou-nan•, p. 260.

Panawataran = Skt. avatära with the Javanese affixes pa- and -an and the long vow:_Is shortened to meet the exigencies of the meter; signifying •the place of an ?Vatara•. Cp. the Chinese ch.iang (11) = •to descend", •to send down • I which occurs m both parts of the Chinese text. c . _DE CASPARis, Prasastl Indonesia, vol. 2, pp. 323-324, slightly modified. A trans­nption of the original text appears on pp. 313--314.

(31) ~~

103

during the Churning of the Ocean 31• The inscription, whose provenance is unknown, was erected by Dyah Lokapäla 32, whose regnal namewas Rakai Kayuwani, in 856 during a Saivite renaissance. Despite the difference of nearly four centuries in the dates of the Chinese and the Javanese texts, the two passages seem to be remarkably similar in spirit: in both a luxuriant growth of vegetation is held to signify that a particular locality is an appro­priate setting for a Saivite theophany 33 • In view of this similarity I am inclined, without any great degree of conviction, to prefer the first of the translations proposed above.

The outstanding event in Cambodian history in whidl Siva figured on a mountain occurred in 802, when J ayavarman II signified the re-unification of the kingdom, its political autonomy, and the institution of an official ~ult focussed on hirnself as cakravartin by raising, according to the ritual of the sacred Vinäsikha, a liiiga that subsequently in Khmer epigraphy was accorded the honorific '~Lord of the Universe who is King" (Skt. devariija; Old Khmer Kamrateil jagat ta räja)34• The establishment (sthiipana) of the liilga took place on Mount Mahendra (Mahendraparvata), whidl has long been identified with the Phnorp. Kulen, the sandstone plateau that overlooks the Ailkor plain from the north 35• At first glance is would seem reasonable to recognize devariija as a designation of Indra, who is commonly charac­terized as '~King of the Gods", and who might be expected to dwell on the "Mount of the Great Indra '~. But neither have cults sudl as that allegedly instituted by Jayavarman II ever been associated with Indra, nor is the liizga a symbol of that god. The place of this cult in the congeries of evolving beliefsthat we call Hinduism becomes apparent only when we turn to texts from the subcontinent itself.

31 Cf. Abhidharmakosa, lU (p. 161 of the translation by L. DE LA VALLEE-PoussiN), composed by Vasubandhu in the 5th century A.D.: DE CAsPARIS, Prasasti Indonesia, vol. 2, p. 304, note 94.

31 Louis-Charles DAMAlS, "Epigraphische aantekeningen I: Lokapala-Kayuwangi", Tijdschrift voor Indische Taal-, Land- en Volkenkunde uitgegeven door het Bata­viaasch Genootschap van Kunsten en Wetenschappen, deel83 (1949), pp. 1-6.

33 For the significance of such hierophanies in a wider context see, int. al., Mircea ELIADE, Le mythe de l'eternel retour: archetypes et repetition (Paris, 1949).

u George CoEots and Pierre DuPoNT, "Les steles de Sdök Käk Th01p, Phnoyp Sandak et Prah Vihär", Bulletin de l'Ecole Francalse d'Extreme-Orient, tome 43 (1943- 6), pp. 83 and 103-104. Cf. also CoEn:Es, "Les expressions Vr~ Kamrateil aii et Kamrateil jagat en vieux-khmer•, The Adyar Library Bulletin, vol. 25 (1961), PP· 447-460. Cf. also Hubert DE MESTIER ou BouRG, • A propos du culte du dieu-roi (devaraja) au Cambodge•, Cahiers d'Histoire Mondiale, tome 11, 3 (1968), pp. 499-516.

115 Etienne AYMONIER, Le Cambodge, tome 1 (Paris, 1900), p. 428; George CoEo:Es, "Les capitales de Jayavarman u•, Bulletin de l'Bcole Francalse d'Extreme-Orient, tome 28 (1928) , pp. 117-119; Philippe STERN, "Travaux executes au Phnom Kulen• , loc. cit., tome 38 (1938), pp. 151-173. Recently George CoEots has demonstrated that the use of devaräja and kamrateil jagat ta räja with the connotation stated here was an anachronism in a ritual ascribed to 802. Most probably the terms were insinuated into the Sdok Käk Thom inscription in 1052, some two and a half centuries after t_he event, by a member of the brähmat;t gotra whose fortunes had become linked w1th that particular cult. It is CoEnts's conte.ntion that the terms in question had been devised as late as the reign of Jayavarman IV (921-941), a monarch of dubious legitimacy, as a means of validating his own claim to the throne. Henceforward, the essence of kingship would inhere not in the person of the ruler but in the abstract notion of royalty, of the kingdom at large [•Le culte de Ia royaute divine au Cam-

104

Some ofthe relevant texts are in Sanskrit. The Ka.Syapasilpa 36, for instance, the Tantric Nisväsatathasamhitä 31 , the Svacchandatantra 38, and several Ganga inscriptions ranging in time from the 6th (or perhaps 7th) century to the 13th or 14th 39, all refer to Siva, not Indra, as residing on the Mahendra­parvata. And always Siva is "the god of the mountain". As Siva Girisa he is the god who slumbers upon the mountain; as Siva Girisa he is the Iord of the mountain; as Siva Giritra he is the protector of the mountain; and his consort is Pärvati or Haimavati, goddess of the mountain, or Sikharavä­sini, she who resides on the peaks 40• But it is still the Sanskrit name Mahen­dra that is associated with the god, and this is a form that cannot have been the original of the Chinese transcription Mua-tgm. Somewhat unexpectedly in the general context of the brähmai).ization of Southeast Asia as it has been customarily presented, it is in Tamil (not Sanskrit) Saivite Iiterature that the source of the Chinese transcription must be sought. The relevant texts have recently been assembled in convenient translation by Professor Filliozat 41 ,

from whose versions the following citations are taken. The earliest extant Tamil reference to Siva as the Kingof the Gods occurs

in the 6th-century Tiruvirattaima.I,limälai of the woman saint Käraikkä­lammaiyär, whounequivocally designates thegod as "Lord of thelmmortals" (amarar piraN) 43• But the most numerous and most explicit passages en­sconcing Siva on Mount Mahendra are to be found in the Tiruväcakam, the "Sacred Utterances" that constitute a veritable spiritual autobiography of the Tamil saint MäiJ.ikkaväcakar, perhaps the greatest of all exponents of the Saivasiddhänta, who lived probably during the 9th century:

bodge", in Himansu Bhusan SARKAR (ed.), R. C. Majumdar Felicitation Volume (Cal­cutta, 1970), pp. 56-62 and 65--66]. Of course, this re-interpretation of the devel­opment chronology of the cult of the devaräja in no way contraverts the fact that Jayavarman II did establish a Saivite cult on the Phnom Kulen in 802 in the manner of his predecessors. ·

38 Anandäsrama Sanskrit Series, 95, 1, 1. 37 Nisväsatathasamhitä, 42a. Cf. Prabodh Chandra BAGCm, Studies in the Tantras

(Calcutta, 1939), p. 95. 38 Kashmir State Survey, 48, Patala I. Cf. D. C. SmcAR, "The Saktas Pithas", Journal

of the Asiatic Society of Bengal, Letters, vol. 14, no. 1 (1948), p. 90. 39 E. G., F. KmLHORN, "Chikkulla plates of Vikramendravarman 11", in E. HULTZSCH

(ed.), Epigraphia Indica, vol. 4 (1896--7), pp. 197-198; R. C. MAJUMDAR, "Pherava grant of Samantavarman, King of Kalinga, year 185", in Joc. eil., vol. 27 (1947-8), p. 1.13; J.F. FLEET, "Sanskrit and Old-Canarese inscriptions", in J. Burgess (ed.), The Indian Antiquary, vol. 13 (1884), pp. 120 and 123. Other references not available to ~e ~r~ cited by FILLIOZAT, •New researches on the relations between India and Cam-04~1a , lndica, vol. 3, no. 2 (1966), p. 102. u BHA'ITACHARYA, Les religions brahmaniques, p. 22.

41 J~an ~ILLIOZAT, "New researches•, pp. 101-102.

u TJIUVIIattaimru:timälai, 13: cited from FILLIOZAT, Joc. cit., p. 102. . Cp. also XLIII, 2 where Vi$t;tU, Brähmat;t, and Indra come to pay homage to

Siva, .who bestows on them his grace. Incidentally, the Saiva-Siddänta accords high prommence to the operation of divine grace (Tamil arul), and it is a moot point wh~t?er the word en {32), a Chinese near-synonym, was employed in Nägasena's ~et~tton . (p. 98) as an unconscious expression of bis Buddhist sympathies or eliberately as an accurate rendering of a cardinal Saivite concept.

(32) -~·

105

11, 8-10. Taking as his abode the great mountain Mayentiram [the Tamil vocalization of Sanskrit Mahend.ra] where he establish­

d bis seat and his glory, there he granted the grace of mani­festing the tradition ( ägamam).

II, 19-20. From his seat on the Mayentiram, he [Siva] granted the grace of giving them [the Ägamas] with his five fruitful mounths.

II, 100. He [Siva] who possesses the great mountain of the mantras, the Mayentiram.

Furthermore, Siva was Kingof the gods: XIX, 7. For Indra, Vi~I)U, Brähma:r;t and the other gods, he is the King.

Brähmal) and Vi~I)u call him: "Our great one, the Lord of gods ( ernperumaN RevarpiraN)."

XXVIII, 9. He is invoked: "0 God of the gods themselves." XXXVII, 7. 0 King of those who are above. The pre-eminence of Siva among the gods is also illustrated in the following passages: VIII, 3. Indra, Vi~Qu, Aja and the other celestials are beyond [this

world]; Siva has been granted the grace to come to the Earth ...

VIII, 4. The gods who went to the sky, Vi~:r:tu, Aja and Indra, have dried themselves in the jungle [in performing tapas], rising like ant-hills and hardly do they see him [Siva] 43 •••

IX, 3. He is the king of those who are beyond [this world]. More generally according to Mä:r:tikkaväcakar, Siva is the luminous pro­

tector of the Mayentiram [Mahendra] (XLIII, 9), the god of the Kailäsa (II, 146; VI, 34, 40; XXIII, 10), and-in Filliozat's summary phrases- "the King and owner of every holy place, of every town, of every kingdom, of the universe, and specially the king of the Gods" 44 •

Phonology and context here combine to support the conjecture that in Mayentiram, the Tamil form of the name for the abode of Siva, is to be discerned the origin of the Chinese Muä-t<}m. Another passage in the Tiru­väcakam goes a long way towards confirming this interpretation. In Hymn XLIII, 2 in particular, and generally in numerous subsequent formulations, Siva is said to have taken the form of a bunter on the Mayentiram. Now it is weH known that, according to later inscriptions the capital of • B'iu-n<}m, situated at the foot of the cosmic axis mundi regnique, at one time was Vyädhapura 45 , hitherto usually translated as the City of the Hunters, but now surely to be rephrased, as the rules for Sanskrit compounds permit, as theCity of theHunter (Siva). It is perhaps significant that theHsin r'ang-Shu

u FILLIOZAT, •New researches•, p. 101. The precise location of the Tamil Mayen­tiram is uncertain, but Kalidäsa in his Raghuvarhsa clearly assigns it to the An~ra country, while the modern commentator Tal)turapayar tentatively identifies it w1th the mountain of Tiruppati (cited by FILLIOZAT, Joc. cit., p. 102]. .

45 CoEots, •tes traditions genealogiques•, p. 127. On a stele from Vät Cakret .m the near vicinity of Ba Phno~ (lnscriptions Sanskrites de Campä et du Ca.mbOdfle: no. LXIII) the Sanskrit honorific of the capital is combined with adri = •mounta~ in the Saivitic phrase Adrivyädhapuresa = Siva of Vyädhapura-on-the-Mounta.m. Cf. also VnQl11 Vyädhapura mentioned in an inscription from Kok Ros~i (CoEDES, Inscriptions du Cambodge, vol. 6 (1954), pp. 17~179).

106

preserves the Old Khmer translation of this honorific, dmak or dalmak, und er the transcription • d'ak-mjuk (ss), but makes no reference to the Sanskrit form 46. The Tang history, although dealing officially with the period between 618 and906, is here describing the situation in • B'iu-ngm prior to the abandonment of Vyädhapura as a result of military pressure from the north­ward during the second half of the 6th century. lt is, therefore, not unlikely that Vyädhapura, the City of the Hunter, may have been the capital some three-quarters of a century earlier when Nägasena was reporting that Siva, in Saivasiddhänta Iore the Hunter, descended on the Mayentiram. There is no epigraphic record of the Saivasiddhänta as such in either India or Cam­bodia prior to a Pallava inscription of the 7th century from Käiicipuram 47,

but the PuRanäNuRu, the oldest collection of Tamil verse extant, attests that the cult of Siva, in association with Vedic recitations and sacrifice, was popular in South India in the time of the Pä1;1Q.yan King Palyäkacalai, who is thought to have reigned prior to A.D. 400 48•

If, as we have tried to show, Siva the Hunter on the Mahendraparvata was recorded in • B'iu-ngm in 484 49 , and if Saivite cults existed in the Tamil country before 400, might they not have been introduced into Southeast Asia as early as the 3rd century A.D. - or possibly even earlier - and might not the sacred mountain, the cosmic omphalos from which either the ruling dynasty or the territory (or both} took its name have been ab initio the Mayentiram?

The use of the Tamil form of a name in a deposition submitted to the Chinese court in 484 is at first sight surprising in view of the general function of Sanskrit in the early centuries of the Christian era as the language of literary communication both within the Indian subcontinent and abroad, but it is not the only instance of Tamil cultural influence in southern Indoch.ina during the • B'iu-ngm period, _nor is it the earliest. In the style of an ancestor

48 Hsin T'ang-Shu, 222, 4a-b. The significance of the transcription • d'ak-miuk

was first recognized by George CoEn:Es, Inscriptions du Cambodge, tome 2 (Hanoi, 1942), p. 110, note 5. Cf. also "Etudes cambodgiennes, XXXVI: quelques precisions s(ur la fin du Fou-nan", Bulletin de l'Ecole Francaise d'Extreme-Orient, tome 43 1943----{>), p. 4.

Some years ago Professor 0. W. WoLTEBS drew attention to a passage in the 3rd­~entulJ Fu-nan t'u-su of K'ang Tai [34], now preserved only in the T'ai-p'ing Yil-lan, m Whl<h a king of • B'iu-ndm is said to have captured and tamed mighty elephants as a means of inducing the submission of neighboring kingdoms [WoLTERS, private CO~unication cited in CoEDES, Les etats hindouises d'Indochlne et d'Indonesie r~s, 1.964), p. 75, note 1]. The implication might have been that the King, by identi-Ying wtth, and thereby emulating, Siva as the divine Hunter on the Mahendrapar­

yata, and perhaps also as Pasupati, the Lord of Beasts, was able to subdue his rivals m.the same way as the god hunted and tamed the wild an.imals on his mountain.

7 Fn.uoZAT, •New researches" p 100

48 Ibid. , . .

•• It is weil known that in their transcriptions of fo.reign names medieval Chinese authors often attempted, with varying degrees of success, torender simultaneously s~und ~nd. sense. It is just possible - though I advance the idea with a good deal ~ trep1dat1on - that the form • Mud-t~m was intended to convey not only the name

ayentiram but also the sense of •[the place where] Ma[hesvara] lingers•.

(33) ~!& C34J •~: ~m±ft:t

107

of a ruler tributary to • B'iu-n9m who is mentioned on the famous stele from Vo-c~nh Filliozat has discerned a Tamil royal title 50• The ancestor in question appears on the stele as Sn Mära, which Filliozat has shown, in the context established by the inscription, can only have been a Sanskrit render­ing of MäRaN, a frequent element in the titularies of the Pä:Q.<;lyan kings of Maturai. After a scholarly debate Iasting more than eighty years the date of the Vo-c~nh inscription now seems to have been firmly ascribed on palaeographic grounds to the 3rd century A.D. 51 • In South lndia the title mäRaN is found not long after the beginning of the Christian era, occurring for example in the PuRanäNüRu 55.6, 56.21, 57.3, and 198.27. But Mära or MäRaN was an ancestor, probably of the fourth generation, of the ruler who caused the inscription to be cut in the third century, which would probably place bis rule in the 2nd century A.D. However, as Filliozat has shrewdly observed, ''On peut aussi se demander si le roi Srimära a jamais regne precisement la ou son descendant a fait graver la stele et s'il ne s'agit pas d'un roi Pät;t<;lya, comme son titre de MäRaN semblerait l'indiquer, dont ce descendant evoquerait dans un royaume d'Indochine le Souvenir et la tradi­tion päl)Qya" 52• The point of concern in the present context is that in the 3rd century a Tamil name, for one reason or another, was important enough to be recorded on stone.

None of the foregoing throws light on the location, as opposed to the iden­tification, of the Mayentiram in • B'iu-ngm. My own preference is for Coedes's identification with the Vra~ VnaiTl the Ba-PhnoiTl of modern times (p. 98 above), though against this equation can be adduced an apparent ab­sence of definitive archaeological confirmation on aerial photographs 53• The vicinity of the hill has yielded no inscription earlier than the reign of isäna­varman I, who probably came to the throne in about 615. But this Iack of material evidence cannot be regarded as conclusive until thorough archaeo­logical investigation has been undertaken on the ground in the neighborhood

50 Jean FILUOZAT, •t'inscription dite de Vo-c~h·, Bulletin de J'Ecole Francalse d'Bxtreme-Orient, tome 55 (1969), pp. 107-116.

51 The Vo-c~h stele, on this dating, bears the earliest extant Sanskrit inscription not only in Indochina but in the whole of Southeast Asia. The controversy as to its date may be followed in: BARTHand BERGAIGNE, Inscriptions sanscrites de Campä et du Cambodge, p. 191; Louis FINOT, •Notes d'epigraphie•, Bulletin de J'Ecole Francalse d'Bxtreme-Orient, tome 15 (1915), p. 3, and •seance du 14 janvier~ , Journal Asiatique, tome 210 (1927), p. 186; D. C. SmcAR, •oate of the earliest Sansknt inscription of Campa•, Journal of the Greater India Society, vol. 6 (1939), pp. 53-55, and "Date of the earliest Sanskrit inscription of Campa•, The Indian Historical Quarterly, vol. 17 (1941), pp. 107-110; George CoEDES, •The date of the Sanskrit inscription of Vo-c~h·, The Indian Historical Quarterly, vol. 16 (1940), pp. 484-488 ; Emile GASPARDONE, •ta plus ancienne inscription d'Indochine•, Journal Asiatiqu~, tome 241 (1953), pp. 477-485, and •t'inscription de Vö-canh et les debuts du sansknt en Indodline•, Sinologica, vol. 8, no. 3 (1965), pp. 129-t36; K. Kumar SARJCAR, •The earliest inscription of lndochina • , Sino-Indian Studies, vol. 5 (1956), pp. 77-87; Kamalesvar BHATTACHARYA, •Precisions sur Ia paleographie de l'inscription dite de yo-canh•, Artibus Asiae, vol. 24 (1961), pp. 219-224. Cf. also Claude JACQ~es,

Notes sur la stele de Vo-c~·. Bulletin de J'Ecole Francalse d'Extr~me-Ouent, tome 55 (1969), pp. 117-124.

51 FILLIOZAT, •t'inscription•, p. 116. 53 Lo~s MALLERET, L'ardleologie du delta du Mekong, vol. 3: La culture du Fou­

nan (Pans, 1962) , p. 453.

108

of the Ba-PhnOip. If an alternative hill had to be sought as the original of Nägasena's Mayentiram, then the Phnmp Da overlooking the city known today as AD.kor B6rei, would suggest itself. Here an extensive rampart, vestiges of a cult centre, and other ardlaeological remains bear witness to the former existence of an important settlement, sufficiently important in fact to warrantnot only the canalization of the Chäu-döc river leading north­westward from the Bassac, but also the cutting of an independent dlannel linking AD.kor B6rei to the port at Oc-eo and ultimately to the Gulf of Siam. One difficulty with this argument would be the fact that, as the history of • B'iu-n9m is currently understood, Ankor B6rei became the capital following the abandonment of Vyädhapura 54 • Although the presently available evi­dence, mostly inferential in dlaracter, would seem to favour the Ba-Phnolll as the most likely locale for the Southeast Asian Mayentiram, this is a matter on whidl definitive certainly will have to await the advent of new evidence, either ardlaeological or epigraphic.

fi 54

George CoEots, •Etudes cambodgiennes, XXXVI: quelques pnkisions sur Ia 0 du Fou-nan•, Bulletin de J'Ecole Francaise d'Extreme-Orient, tome 43 (1943--6),

p. ~· A third alternative recently proposed by Jean BoxssELIER, namely that the ~api.tal of • B'iu-n9m was Iocated at the site of present-day U-Thong in Central ~ailand, does not seem to me to be a viable proposition in the light of present

f wledge: ·u T'ong et son importance pour l'histoire de Thailande" , Silpakon, Vo · 9 (1965), pp. 27-30; •L'art de Dvaravati•, Silpakon, vol. 11 (1968) , PP· 34-56.

109