29

Some Notes on the EpigraphicHeritage of Srivijaya

By Prof. Dr. J.G. de Casparis

Inscriptions associated with the Srivijaya, though

only a few, serve as an important source of the

empire's history.

Compared with the considerablenumbers of Old Javanese and OldBalinese inscriptions, those found inthe island of Sumatra are relativelyfew. Of those available, the inscrip-tions of the fourteenth-century'king Adityavarman constitute themajority; a relatively small numbercan be associated with Sriwijaya.

Only seven inscriptions actuallymention the name of Sriwijaya,while a few others, mostly frag-mentary, can on various groundsbe attributed to this great empire.Of these seven inscriptions, three,viz. those of Kedukan Bukit, Tela-ga Batu' (in fact from Saboking-king, east of the city of Palembang)and Talang Tuwo belong to thecity of Palembang or its immediatevicinity. The inscriptions of KarangBrahi and Palas Pasemah belong tothe western part of the province ofJambi and the southernmost partof the province of Lampung res-pectively. The sixth, the inscrip-tion of Kota Kapur, comes fromthe island of Bangka and the lastone, probably from 'Ligor' (Na-khon Si Thammarat) in peninsularThailand.

The inscriptions dated have been

Prof. Dr. J. G. de Casparis is a professor at the institute Kern, Rijksuniversiteit le Leiden in Nederland.

written towards the end of theseventh century: 683, 684 and686. The Ligor stone, dated 775A.D., is considerably later, but theyall fall within a narrow period ofless than a century. The undated,mostly fragmentary inscriptionscan, mainly on paleographicalgrounds, be assigned to approxi-mately the same period. The lan-guage used in these inscriptions, ismostly old.Malay; six of the sevenwere written in that language. TheLigor stone stands somewhat apartfor it is drafted in Sanskrit.

It is therefore clear that theepigraphic heritage of Sriwijaya be-longs to the first century of theexistence of the state and this limitsits usefulness. Despite such limita-tions, the inscriptions are, however,of immense value as they consti-tute the only authentic source, i.e.written source, of the history ofthis great empire. The Chinese,Arabic and Indian sources are, ofcourse, of immense importance, butthey have to be used with greatcare. The Chinese sources deal onlywith certain aspects of Sriwijayaand may sometimes misunderstandthe actual institutions. On the otherhand, the Arabic and Indian ac-counts usually suffer from grossexaggerations.

The inscriptions supply us withreliable information if — and thatis an important limitation — onecan read and understand them pro-

perly. If one considers that theAsokan inscriptions, despite inten-sive study for more than 150 years,still pose problems that are onlygradually being solved, it is hardlysurprising that the Sriwijayan in-scriptions, which have been serious-ly studied for less than half theabove-mentioned period, still pre-.,sent serious problems of readingand interpretation. Only a fewpoints are discussed in this paper.

The paleography of the Sriwijayan

inscriptions

The inscriptions dated betweenA.D. 683 and 775 are written inscripts usually called 'Pallava'duetoits similarity with the script of thePal lava charters of about the sameperiod. During this time, a broadsimilarity between the scripts usedin southern India, Sri Lanka andparts of mainland and insularSoutheast Asia was observed. Thereis not yet any clear local differen-tiation. Hence, in most cases, itwould be very difficult, if not im-possible, to assign an inscription ofunknown origin to a particular areaon the basis of paleographyalone.

As to the Sriwijayan inscriptions,one can clearly distinguish threedifferent types of script amongthose issued between 683 and 686.

The inscription of Kota Kapur(island of Bangka) is written in a

30

Map of the sites of the inscriptions

very regular and elegant type ofscript. It has relatively long andslender letters with some decorativeelements which do not interferewith the basic function of thescript, i.e. unambiguous communi-cation. The inscriptions of KarangBrahi, Telaga Batu and some of thefragmentary texts are also drafted

in this calligraphic but functionalscript. This style of writing differsconsiderably from that used in theinscriptions of Mulawarman in EastKalimantan, Purnawarman in WestJava and Mahendravarman in Cam-bodia which can be dated back tothe beginning of the seventh cen-tury.

The main difference rests onwnat I have elsewhere called the'equal height' principle. All simpleletters, excluding vowel marks, liga-tures, viramas etc., have the sameheight in the Srivijayan inscriptionsearlier mentioned. This manner ofwriting facilitates reading for itimmediately separates the basicforms from their accessories.

The pre-Sriwijayan, so-called Pal-lava, inscriptions including, e.g. theMahendravarman inscriptions, showa number of letters such as a", ra, la and ka which extend far below orabove the frame within which mostof the other letters are contained.The ka, for example, has a kind ofdownward extension which in otherletters would be interpreted as a subscript -u. From this point ofview, the type of script used in theSriwijayan inscriptions must be re-garded as an improvement as itpromotes both legibility and order.

It is difficult to determine wheth-er this innovation was first used inSriwijaya and subsequently follow-ed by others. One finds other exam-ples of scripts using the equalheight principle in mainland South-east Asia in about the same period,i.e. the last quarter of the seventhcentury.

A totally different type of script,though also adhering to the equal-height principle, is used in theKedukan Bukit inscription. Thescript has no trace of ornamenta-tion. It is purely functional, squareand straight. In this respect, it clear-ly foreshadows later developmentexemplified by the Ligor inscrip-tion and the Javanese script of in-scriptions of A.D. 760 and later.

Finally, a third type of script,exemplified by the Talang Tuwoinscription, occupies a sort of in-termediate position. Its script isstylized too, but not to the extentthe Kota Kapur inscription is. It isslightly decorated and not as slen-der as the Kota Kapur script. Let-ters such as the pa, sa ma, ga and a few othes are just about as broad asthey are high.

No Chronological Significance

However, it should be stressedthat the difference among the

31

scripts is only stylistic. It has nochronological implications. Hence itis not surprising to see quite a num-ber of different styles of writingused side by side in the short in-scriptions at Mahabalipuram, allgiving different birudas and epithetsof King Rajendravarman.

In the case of the Sriwijayan in-scriptions, it is striking that themost simple, 'ordinary', type ofscript is used in the Kedukan Bukitinscription. It provides the authorone good reason to reconsider thecontents of this important, but'difficult, inscription.

The Kedukan Bukit Inscription

Whereas most other Sriwijayan in-scriptions, such as the stone inscrip-tions of Kota Kapur, Talang Tuwo,Telaga Batu etc., are engraved incarefully cut stones with smoothsurfaces, the Kedukan Bukit in-scription is cut in an ordinary boul-der of the kind used for thenumerous siddhayatra inscriptions.The simplicity of the stone and ofthe script used for the text is inagreement with the contents of theinscription.

This is by no means a prasasti of the ordinary kind (such as onesees, for instance, in the Ligor in-scription of 775), but a rather dryand factual account of the move-ments of the dapunta hiyang, per-haps the king or else, a high officialof the empire.

If it is a report, the next questionthat arises is: by whom and forwhom is it written?. Coedes was,of course, clearly aware of thisproblem, for he started the dis-cussion of the inscription with thewords: quelle est au juste I'objet de cette inscription?' His reply is,as the learned scholar himself ad-mits, somewhat unsatisfactory andmainly negative in that he mainlycombats Van Ronkel's interpreta-tion without giving a clear alter-native.

The most striking aspect of theinscription is its reference to sid-dhayatra both at the beginning andat the end. It started with theDapunta Hiyang embarking (naik samwau) in order to seek (literally:take, manambil) siddhayatra, andend with him making Sriwijaya

jayasiddhayatra, prosperous and ...(here the stone is broken off).

In this connection, it should benoted that there is a strange lapse inthe otherwise excellent transcrip-tion by Coedes. It has apparentlycrept into the transcriptions of laterscholars, viz. the omission of jaya following sriwijaya, presumably bya kind of haplology, not rarelyfound in manuscripts.

One may wonder whether themeaning of jayasiddhayatra is verydifferent from that of siddhayatra itself when jaya, in its presentusage, merely adds a kind of vic-torious or honorific tinge to somenames, for example in Jaya Ancoland Irian Jaya which used to be justAncol and Irian Berat. On theother hand, one may argue thatthere must have been a clear dif-

air). However, two objections wereraised against this view. First, whyis siddhayatra, not siddhiyatra used.The second, more serious, objectionstems from the Mahayanicdoctrinewhich states that such supernaturalpowers can be acquired by a personbut can hardly be conferred uponan empire.

The interpretation proposed byB. Ch. Chhabra, who discussed theterm in connect ion with the inscrip-tion on the stone slab of an engrav-ed stupa from Bukit Meriam, ismore acceptable to the author. Inthis inscription by the captain ofa ship named Buddhagupta, hewished that 'they all' (presumablyall the inmates of the ship) may besiddhayatrah in all their enterprise'.Chhabra is right that there canhardly be any question of magical

The Kedukan Bukit Inscription is a factual account

of the movements of Dapunta Hiyang.

ference between the two.

It is, however, difficult to cometo any conclusion without firstgetting a clear idea of the meaningof siddhayatra, This term has givenrise to much discussion. VanRonkel took it as a synonym of thelater term berkat, derived from Ara-bic barakat, a blessing that a holyor merely older person can give tosomeone on the point of makingan important decision, such as a marriage, a voyage or perilous en-terprise. This meaning fits well thecontext of the Kedukan Bukit in-scription. The 'seeking' of siddha-yatra clearly preceded an importantenterprise by the Dapunta Hiyangwho, some four weeks later, wouldleave (the capital?) with a considera-(

ble number of troops. After his re-'turn, he would have conferred thisnewly acquired blessing on hiscountry, i.e. Sriwijaya.

Coedes prefers a different inter-pretation and compares siddhayatrawith the well known word siddhi which is usually an indication ofsupernatural powers of the Buddhaor of Bodhisattvas of the higheststate (such as flying through the

powers in this case. The meaning'success' or 'successful' seems to berequired by the context.

Such an interpretation wouldalso fit the context of the KedukanBukit inscription. The DapuntaHiyang would, at the first mention-ed date, have left (the capital) toseek, by prayers and meritoriousdeeds, success in the enterprise,probably a military campaign,which he was to undertake fourweeks later. It has been suggestedfirst by Mr. Boechari that the dateof his departure, the 11th of Vaisa-kha, was chosen in order that themeritorious acts (perhaps of thekind mentioned in detail in theTalang Tuwo inscription of 684)would be proclaimed on the 'ThriceBlessed Day' of all Buddhists: theFull Moon of Vaisakha. This wouldimply a two- or three-days' journey,assuming that the Dapunta Hiyangwould have made sure to be at hisdestination well in advance of thescheduled date of his pious acts.After the, presumably successful,completion of the enterprise orcampaign, the Dapunta Hiyangwould have decided to share the

32

success of his victory with his sub-jects, probably in the form ofbooty and captives.

The site of the siddhayatra cere-mony, no doubt, should be iden-tified as Telaga Batu', the StonePond at or near Sabokingking,where a considerable number ofstones with the inscription siddha-yatra or jayasiddhayatra have beenfound. Where Dapunta Hiyangcame from is, however, more dif-ficult to determine. One might sug-gest a place near Kedukan Bukit,the presumed site where the inscrip-tion was found, but the journeyfrom there to Telaga Batu' wouldonly take a few hours.

The author doubts, however,that Kedukan Bukit is really theplace of origin of the stone inscrip-tion of 685. The stone was dis-

which the compaign was directed.Coedes' reading mata jap is qualifi-ed as uncertain and does not makesense. Krom suggested correcting itto malayu, which does make sensebut does not correspond with thestill visible parts of the aksara. Inparticular, thesecond letterdoesnotappear to have been la, but seems tohave been either ka or to. The thirdletter, of which only the upper partremains, could well have been a ya, but the vowel mark for subscript-u is conjectural as the lower part ofthe letter is broken off.

In his paper submitted for theSPAFA Workshop, Mr. Boecharimade an ingenious conjecture con-cerning the reading of this passage.He noticed traces of the beginningof a subscript -u below the ma, which is clearly visible. He there

Dapunta Hiyang's embarkation point and destination

are still an object of research.

covered in 1920 in the words ofCoedes 'chez une famille malaise de Kedukan Bukit, au bord du Sungei Tatang, affluent du Musi, au pied de la colline de Bukit Seguntang. 'Coedes adds in a foot-note that a 'resident' of Palembangat that time, L.C. Westenenk, ex-plained that the stone had, for a long time, been in the possession ofthe family which used it as a mas-cot during boat races. As the storyof how the family came to possessthis inscribed stone is unknown, itis difficult to ascertain its true siteof origin. It could be Telaga Batu'as a stone fragment with part of thesame text, now in the Jakartamuseum, originates from that site.In that case, one has to look for theresidence of the Dapunta Hiyang,no doubt also the capital of thekingdom, at some place a few days'travel by prau away from TelagaBatu'.

Where Did Dapunta Hiyang Go?

Different suggestions have beenmade regarding the place against

fore, read the first two syllables asmuka or mukha. He also pointedout that the left part of the. letterread as ja or ya (or jap by Coedisand yu by Krom) could well be theupper part of an initial u-. He agreedwith Coedes that what followedmay be the upper part of a pa, buthe saw no virama above it thoughhe recognized an anusvara. Thiswould read upang in front of which(di muka) the Dapunta Hiyangwould have arrived. Indeed, animportant village called Upangdoes exist at some 30 km. distancedownstreams of Palembang. Itsmain part consists of a nearly fourkm. long 'boulevard' entirely builton poles and planks above the coast.Was this indeed the place againstwhich the efforts of the DapuntaHiyang were directed?

Having examined the passage onthe original at the Musium Pusat inJakarta, the author believes that thereading, though by no means im-possible, is hardly likely. The twovertical strokes which were identi-fied as the upper part of an initialu= and the left downward stroke ofa pa seem to be too close together

to belong to separate aksaras. Theymost probably have to be takentogether to form a ya with or with-out -u. Boechari's reading of mu instead of ma is, a definite improve-ment. However, too many possibili-ties exist, so the best decision is towait for further discoveries.

Another Passage NeedingInterpretation

This is not the only uncertainpassage in the Kedukan Bukit in-scription. In lines 9 and 10, a smalllacuna is present between marwuat wanua at the end of line 9,.andsriwijaya jayasiddhayatra , at thebeginning of line 10. Coedes tran-slates this as: 'vint et fit le pays Srivijaya, doue de puissance magi-que', but his interpretation seemsto have ignored the lacuna. Nobodyknows what the Dapunta Hiyangmade or did (marwuat) in order thatSriwijaya would become successfuland prosperous. The parallel versiondiscussed in Pras. Indon. I I , I e isof no help in this case, as the gapis even larger at this spot.

The account of the mysteriousmovements of the Dapunta Hiyangin the spring of A.D. 683 is still a problem as it was when Coedesstudied it half a century ago. Ananalysis of the problems of read-ing and interpreting the inscriptionmay pave the way to their ultimatesolution. However, no progress canbe made until new materials arediscovered.

Other Sriwijayan Inscriptions.

The other inscriptions are muchmore straightforward. They relatethe various activities of one or moreof its earliest rulers. As one mayhave expected in such a far-flungmaritime empire, these activitiescentred around punitive expedi-tions (Kota Kapur); curses againstall possible evildoers, especiallypotential traitors (Kota Kapur,Karang Berahi and Palas Pasemah);and also pious foundations (TalangTuwo).

It is perhaps no coincidence thatthe most carefully engraved inscrip-tion is that on the hexagonal•column of Kota Kapur; those of

33

Karang Berahi and Palas Pasemahseem no less carefully engraved,but they are badly worn . Onewould expect that imprecationswere an important element of Sri-wijayan policy, correponding tothe threats and blackmail of mo-dern states. Whether such curseswere real deterrents is, of course,impossible to determine. Even ifthey were not, they could, how-ever, be used as a casus belli to jus-tify Sriwijayan intervention. Thesecurses, the most elaborate versionof which is found on the stone of'Telaga Batu', are of particular in-terest as they throw some light onthe major preoccupations of therulers. All sort of rebellious ac-tivities are mentioned with precisedefinitions as in a code of crimi-nal law (chapter: 'Crimes againstthe State'). Although it may seemthat some of these definitions areof a somewhat theoretical naturein general, they are based on ac-tual experience and practice.

Curses Against Kins

The Telaga Batu' inscription hasthe most elaborate of the impre-cations. Whereas the shorter cur-ses of the Kota Kapur, Karang Bera-.hi and, presumably, Palang Pase-mah inscriptions are mainly direct-ed against possible disloyalty bylocal representatives of central au-thority, viz. the daws appointed bythe king, the Telaga Batu inscrip-tion includes the very highestauthorities below the king — theyuvaraja, pratiyuvaraja (the princeimmediately following the crownprince in rank) and the rajakumaras (the other princes) — among thepossible rebels.

In publishing the incription in1956, the author added that theking might have most to fear fromhis own relatives eager to take hisplace. Mr Boechari strongly disa-greed and thought of such a situa-tion as 'unthinkable' in an Indone-sian keraton. Yet, the unthinkabledoes happen. The Violent quarrelbetween Amangkurat I and his el-dest son, the Pangeran AdipatiAnom (The later Amangkurat II),contributed to the destruction ofthe keraton at Plered in 1677 andthe flight and death of Amangku-

rat I.In the same period, an even

worse quarrel arose between theSultan of Banten and his CrownPrince, who did not hesitate to con-clude an alliance with the Nether-lands East India Company and hadhis father imprisoned — the oldSultan died a few years later in pri-son. The Buddhist chronicles ofCeylon, a pious Buddhist country,abound in intrigues of this kind.Wars between brothers and cousinsare almost common-place. Onegets the impression that manyprinces felt few scruples if they sawa chance to rise to the throne.

One thing seems certain: thestone containing the most elaborateversion of imprecations, which wasto be 'drunk' by the appointedgovernors (datu) before leaving for

Since the Kota Kapur, KarangBerahi and, presumably, Palas Pase-mah inscriptions are abbreviated re-plicas of the Telaga Batu stone in-scription, the latter should beslightly earlier than the rest. Thisgives a strong additional argumentin favour of the location of thecapital of Sriwijaya at or nearPalembang, at least in its earlyphase. The ancient harbour of theempire could have been on thenorthern bank of the river nearSaboking king, the place where theofficials took off after the oathritual.

Kind of Scripts Used

To return to the scrip, two styleshave been noted: the ordinary,

Imprecations defining rebellious activities are contained

in some inscriptions.

their respective provinces, musthave stood at not too far a dis-tance from the capital. If subse-quently they committed any of thecrimes mentioned in the inscrip-tion, they would automatically bekilled by the effect of the impre-cation (with some help from theking' spies?). This would imply thatthe site of the stone, Sabokingkingon the eastern fringe of the presentcity of Palembang, would not betoo far from the royal compoundin or near the capital. It also impliesthat the royal compound could nothave been too far away fromSabokingking. Thus this inscriptionstrongly indicates that the capitalwas in the Palembang area duringthe period when the stone was inuse for the oath ceremony. Thistime limitation is of course essen-tial for no proof that this oathwas in use throughout the existenceof the empire is available. Its scriptis of the early type, virtually simi-lar with that used in the KotaKapur stone of A.D. 686, suggest-ing that the 'Telaga Batu' stone isroughly contemporary with the for-mer.

functional script of Kedukan Bukiton the one hand and the elegantand decorative script of the impre-cations on the other. The script ofthe Talang Tuwo inscription of 684stands about midway between thesetwo. It is quite regular and square,obviously engraved with great carebut without the elegant shapes ofthe oath inscriptions. Its contentsare also quite different. They speakof a grant by the Dapunta Hiyangof a park for the benefit of all crea-tures, in particular travellers whowould find there lodging and food.This pious gift is part of a pranidha-na by the Dapunta Hiyang whovowed that all his future acts wouldbe directed towards seeking ultimateenlightenment.

The above analysis shows a kindof correlation between the ex-ternal form, especially the script,and the contents of the Sriwijayaninscriptions.

The Old Malay Language ofthe Inscriptions

The language of the majority of

34

the inscriptions — the main excep-tions are the great 'Ligor' inscrip-tion(s) of 775 and a few fragments— has been called Old Malay whichis clearly the ancestor of classicaland modern Malay and Indonesian.It shows, however, a few particula-rities — the prefixes mar (Malayber) ni- (Malay di-) and maka witha causative meaning (maka gila = Malay menggilakan), as well aswords such as inan, 'Jhis'. Theseare all clearly Sumatranese featurespreserved in languages such asMinangkabau and Batak. OldMalay is clearly a bureaucraticlanguage. Though structurally andin its basic vocabulary Austrone-sian, it contains not only a largepercentage of Sanskrit words,especially technical terms of Bud-dhism and administration, but also

stone of Candi Sewu. Despite someminor difference in language, suchas the use of war instead of mar, there is little doubt that the useof Old Malay in Java reflects director indirect influence from Sri-wijaya.

Such influence also appears in a few Malay words found in Old Ja-vanese inscriptions. Owing to theclose relationship between the twolanguages, such influence is oftenimpossible to prove. For example,is tiga in Old Javanese and krama in modern Javanese a synonym oftelu or tlu due to Malay influenceor does it belong to the commonAustronesian heritage?

However, a few examples seembeyond doubt.

The so-called import inscriptionof the Kedu, an edict of King Bali-

Old Malay, the language used in most inscriptions, is the

foundation of classical and modern Malay and Indonesian.

are also among the divine powersinvolved in the oath. One gets theimpression that all kinds of deities,including those supposed to bepowerful in different regions, wereincluded in those imprecations.

It is therefore not surprising that,from the tenth century, the puzzl-ing name of Waprakeswara (orBaprakaeswara) is found at the be-ginning of the imprecations. Beforethis, the name is only attested inthe fourth-or fifth-century inscrip-tions of king Mulavarman in EastKalimantan. It seems unknown inIndia or Cambodia. Therefore, a strong proof indicates that the useof this name in Java must be at-tributed to a close relationshipbetween East Kalimantan and EastJava in the tenth century. TheJavanese, hearing about (or ex-periencing?) the great power ofWaprakeswara in East Kalimantan,must have decided to incorporatethis god, no doubt regarded as a manifestation of Siwa, into theirpantheon.

some functional elements such astathapi, 'however' (modern tetapi).

The. commercial power of Sriwi-jaya may have spread the use ofthis language to other parts of thearchipelago and may thus have con-tributed to the later growth of theIndonesian language.

Linguistic Influence

The first phase of this develop-ment is probably already noticea-ble in the 8th-9th century, when a small number (seven are at presentknown) of inscriptions in OldMalay were found. All but onehave been recovered from northerncentral Java (Sojomerto south ofPekalongan, Dieng, Gondosuli inthe kabupaten of Temanggung).One was uncovered in southerncentral Java, viz. the foundation

tung dated A.D. 907, shows the useof the well known Malay numeralsatu (spelt satu) in a context wherethe meaning 'one' is clearly requir-ed. Is it a mere coincidence that theking's personal name Balitung isidentical with that of the 'tin is-land', commonly spelt Billiton?Since reasons to suppose that thisking married into the Mataramroyal family exist, relations withthe Malay world can by no meansbe excluded.

Another example from the sameinscription concerns the gods andspirits invoked as protectors of thenew foundation. In connection withNagas and other semi-divine crea-tures, Tandung Luah, a name thatis nearly identical with TandrumLuah, is the spirit twice invoked atthe beginning of the oath insriptions of Srivijaya. He serves as theprincipal protector of the kingdom.

A curious enumeration of higherand lower gods of the Hindu pan-theon and Old Austronesian deities

Unity Among Indonesians

The subject of inter-Indonesianrelations in ancient times is fasci-nating but little is known exceptsuch obvious topics as Javaneseinfluence in Bali or in Banjarmasin.However, Java also was influen-ced by the other parts of the archi-pelago. Such exchanges were ofgreat interest in the later shapingof the Indonesian nation. The well-known lists of foreigners in the in-scriptions at the time of Airlanggaand later show that traders fromthe remote parts of the archipelagowere never included in such lists. Inother words, there was a clear dis-tinction., between Indonesians andnon-Indonesians. Despite conti-nuous rivalry, perhaps, a kind ofsolidarity existed between Indone-sians of many different islands.Although Java no doubt played a great part in these inter-Indonesianrelations, it is quite likely that therole of Srivijaya in these develop-ments has been an important, ifnot a decisive, one.