Introduction
The three articles by Hermann Kulke, Jan Wisseman Christie, and Pierre-Yves Manguin published in this volume are revised versions of papers presented at a panel on early Indonesian state formation during the Association for Asian Studies Meeting at Washington, DC, in 1989. Using different kinds of evidence and different methodologies, these authors share a common interest in how new structures of authority were conceptualized and formalized in early Java and Sumatra and how these structures may have responded to pre-existing patterns of hierarchy. These papers reveal possibilities for analyzing early state formation that suggest diversity both in the Indonesian past and in the epistemological and methodological potentialities of various kinds of evidence. They are published here not out of an antiquarian interest in the distant past but more with the conviction that efforts to analyze the past into coherent narratives or models reveal important aspects of contemporary academic and political culture.
K. W. Taylor
By Hermann Kulke
I
As explained elsewhere in a more theoretical context, state formation in Southeast Asia took place in three consecutive phases, which correspond with chiefdom (or local principality), the early kingdom, and finally the imperial kingdom. [1] The present article organizes and analyzes the epigraphical references to the "city" and the "state" according to these three developmental phases. Part I deals with Indonesia‘s earliest inscriptions of Mulavarman and Pΰrnavarman of the fifth century C.E., which depict the transformation of a chiefdom into an early kingdom. Parts II to IV analyze various developmental stages and structural problems of early kingdoms as exemplified by the inscriptions of early Srivijaya and of Central and Eastern Java from the late seventh to twelfth centuries. The final part deals with the growth of the imperial kingdom under Singasari and Majapahit. The emphasis of the paper is the descriptive analysis of the epigraphical evidence rather than an evaluation of its theoretical implications. [2]
The first step towards state formation that is discernible in inscriptions is the transition from chieftaincy to early kingdom. In Indonesia this transition is illustrated very instructively by the famous Kutei inscriptions of Mulavarman who ruled about 400 C.E. in East Kalimantan. [3] These earliest inscriptions of Indonesia, incised on seven yupa or sacrificial stone posts that bear a strong resemblance to menhirs, contain the unique story of the rise of a local chieftaincy and its transformation into an early kingdom within three generations. [4] The story begins with a local leader, perhaps a lineage elder, with the name Kundunga. His Sanskrit title "Lord of Men" (narendra) may have been conferred upon him only posthumously by his son or grandson. Under his son something important apparently took place, as one of the inscriptions reports that he became the founder of a lineage or "dynasty" (vamsa-karta) and assumed the Sanskrit name Asvavarman. Whatever that may have meant in detail, it appears that Asvavarman was able to raise considerably the status of his lineage (vamsa) within his own clan.
The decisive steps towards the establishment of an early kingdom took place in the third neration under Mulavarman. He assumed the royal title raja, defeated neighboring chiefs (parthiva), and ade them his tributaries (karada). Furthermore, he invited Brahmins "who came hither" (ϊhagata) and celebrated grand rituals at a "most sacred place" (punyatama ksetra) called Vaprakesvara, and had a series of impressive inscriptions incised. [5] The meaning of the unusual name of "Lord (Siva) of the Vapra(ka)" is unclear. But the fact that the place is explicitily described as "most sacred" allows us to infer that it had something to do with Mύlavarman‘s rise to power. Sanskrit vapra has the meaning of either "rampart" and "earth raised as the foundation of a building" or "mound" and "hillock." Thus this "Lord of the Vapra(ka)" may have been associated either with the foundation of Mϋlavarman‘s own "town" (pura), which is mentioned in one of the inscriptions, or with a ritual on a hillock. Although the first meaning cannot be ruled out, the latter is more likely in view of other examples known from Southeast Asia where liήgas of Siva (isvara) were consecrated on a hillock (vapra) in connection with the foundation ritual of a polity. The most famous example is, of course, the establishment of a Sivaliήga as Devaraja on the Mahendra mountain in 802 C.E. by Jayavarman II, the founder of the kingdom of Angkor. [6] More important in the context of early Javanese history, however, is the consecration of a Sivaliήga by Sanjaya on a hillock in Central Java in 732 C.E. [7] In regard to Mύlavarman‘s Vaprakesvara, we may even go a step further and infer that the deity Vaprakesvara may have also been associated with the cult of a deified ancestor, the very "root" of the "dynasty" established by Mϋlavarman‘s father. At least Mϋlavarman‘s name, "protege (varman) of the root (mϋla)," makes such an inference quite likely. Finally, it should be emphasized that the making and establishment of the monumental yΰpa stone inscription by the Brahmins (their authorship is mentioned twice in the inscriptions) in a tribal surrounding must have been particularly impressive and successful in raising the status of Mΰlavarman. The Brahmins "who had come hither" were generously rewarded by Mΰlavarman with land (bhΰmi) and thousands of cows.
The analysis of the epigraphical evidence of these earliest inscriptions of Indonesia allows us to draw the following tentative picture of Mύlavarman‘s new polity. At its center was the pura, which, however, certainly did not yet represent a truly urban settlement. The statement that it was "his own (svaka) pura" makes it more likely that, as in later cases, here, too, pura meant the "residence" or kraton of Mΰlavarman. Not very far away from the pura was the "most sacred place" (perhaps on the "sacred mountain") of Mΰlavarman‘s polity, which might have been associated with a Hinduized ancestor cult. The pura may have been surrounded by the dwelling places and lands of the Brahmins "who came here" and by the places of other members of Mΰlavarman‘s lineage (vamsa). Beyond this nuclear area and its adjoining jungle, similar, though most likely smaller, places of other "landlords" (pάrthiva) and lineage elders and their family members were situated. Some of these little chiefs had been defeated by Mύlavarman and thus become his tributaries. This is all we know about Mΰlavarman‘s polity from his inscription. Apart from the pura, the holy k$etra, the land (bhΰmi) donated to the Brahmins, and the surrounding "landlords" (pάrthiva) donated to the Brahmins, no other evidence is known that allows us to define the polity spatially. Furthermore, it is significant that no "officers" of any political function are mentioned in these inscriptions. In Weberian terminology, Mύlavarman‘s authority was thus a purely patriarchal rule based, most likely, on his patriarchal household and its oikos economy. But there was one significent difference. Mύlavarman‘s court was able to attract Brahmins who certainly acted not only as ritual specialists but also as advisers, thus fulfilling a role that comes already near to Weber‘s extra-patrimonial staff. The fact that Mύlavarman managed to invite (and feed!) Brahmins and to have them perform grand rituals and to compose and incise seven impressive yΰpa inscriptions distinguishes Mύlavarman‘s polity from the many little chieftaincies that surrounded him. However, although he claimed to have defeated them, he still remained one of them, though as primus inter pares; whereas these chiefs were called "Landlord" (pάrthiva) Raja Mΰlavarman was praised in his inscriptions as the "Lord of the Landlords" (pάrthiva-indra.
In this connection a small hint from the earliest inscription of mainland Southeast Asia, the famous Vo-Canh inscription from Central Vietnam, is very instructive. [8] It reports an announcement "beneficial to the people" (prajά) by King Sri Mara that he was willing to share all his property with those dear to him. Among these were his sons, brothers, and relatives explicitly mentioned as having been satisfied by Sri Mara while he sat on the throne in the midst of his kinsmen (svajana). Much has been written about this inscription. On palaeographical grounds, its date is usually given as the second or third centuries C.E. G. Coedes and others assumed an identity of Sri Mara with Fan Shih-man, the great king of Funan in the third century C.E. [9] But to my understanding it is absolutely impossible to interpret this inscription as a document of a great Funanese king who, in this case, would have been ruling over a large kingdom that spread from Funan‘s center in the Mekong Delta up the Annamite coast where the Vo-Canh inscription was found. Instead, the Vo-Canh inscription appears to represent a "state" that has to be equated evolutionarily with Mΰlavarman‘s polity. In the Vo-Canh inscription we meet only the king, his relatives, and the people. He sat amidst his own kinsmen and promised to share his property with them. As in the case of Mΰlavarman‘s inscription, there were no officers of any kind present during the ceremony. Some Brahmins must have been around to compose and to incise the inscription, but they seem to have been even less influential than under Mulavarman as they are not referred to at all. Srϊ Mara‘s "state affairs" were therefore clearly the affairs of the "royal family" (rάjakula) which is mentioned in the inscription.
There is yet another fact that brings the Vo-Canh inscription typologically even nearer to the Kutei inscriptions. In its (otherwise very mutilated) introduction, the Vo-Canh inscription, too, mentions a "first conquest" (prathama vίjaya). As in the case of Mulavarman in Kalimantan, Sri Mara‘s family Qcula) or clan obviously had just undertaken some successful raids against neighboring chiefs. In order to ascertain the loyalty of his (envious?) relatives (in this regard the explicit mention of his brothers is noteworthy), Siϊ Mara felt obliged to assure them of his willingness to share the spoils, and he appealed to future kings to do the same. From a structural point of view, we have thus the same situation as we came across in Eastern Kalimantan after Mulavarman had undertaken his first conquests. In this regard it is important that D. C. Sircar, the renowned Indian epigraphist, stated that "the date (of the Vo-Canh inscription) is not much earlier than the 5th century AD." [10] Sri Mara thus would ave been a contemporary of Mulavarman, facing obviously quite similar problems while xtending his authority beyond his own family or clan territory.
Pΰrnavarman‘s nearly contemporary inscriptions of mid-fifth century West Java depict a icture of a slightly more developed early statehood. [11] Pύrηavarman is praised as the "Lord (Tsvara) of the city (nagara) of Taruma" whose predecessor Pΐnabahu already bore the truly royal title "Foremost King of Kings" (rajadhίmjά). Whereas Pϊnabahu had dug a canal passing beside the "famous city" (purϊ), Pΰrnavarman dug another canal that cut across the "cantonment" (sibira) of his grandfather, who might have been identical with Pϊnabahu. In any case, here, too, three generations are mentioned, although in this case already the grandfather seems to have been able to impose his authority upon other chiefs. Pΰrηavarman continued this policy as he is explicitly praised as conquerer of the "towns of his enemies" (aήnagara). In this connection, too, it is interesting to note that Pΰrnavarman, in spite of his conquest, remained a primus inter pares, ruling over his own nagara as did his enemies (art) in their own nagara. As in the case of Mύlavarman‘s (and Sri Mara‘s) inscriptions, we find no mention of officers or any references pertaining to the spatial aspect of his authority. Therefore we have no idea about the character of Pύrnavarman‘s relations with the "cities" (nagara) of the enemies that he claims to have conquered. There was, however, a remarkable difference between these early kings. Pύrnavarman‘s association with the Hindu deities Visnu and Indra [12] clearly indicates that his court has already come under much stronger Indian influence, perhaps over several generations, than those of Mulavarman and Sri Mara. But in Java, too, this influence had not yet fundamentally changed the nature of the state of Tarumanagara.
However, one aspect of the epigraphical evidence of Pΰrnavarman‘s inscription deserves our attention. Whereas Mύlavarman‘s inscription mentions only once a pura (there is no such mention at all in Siϊ Mara‘s inscription), Pύrnavarman‘s four inscriptions refer to nagara, γurϊ, and sibira. The so-called Ci-Arutan inscription clearly praises Pΰrnavarman as the "Lord of Tarumanagara"; the Jambu Rock inscription mentions Taruma and the defeated arinagara; the Kebon-Kopi Rock inscription praises the "Lord of Taruma" (without mentioning nagara)‘, and the Tugu Stone inscription describes the two canals in connection with the "famous purϊ" and the sibira.. The meager epigraphical evidence does not allow a clear distinction between these three terms. But we may infer that sibira. meant a fortified place, perhaps the kraton of Pΰrnavarman or his grandfather, which was situated within the purϊ. This expression then would refer to Taruma as a semi-urban or urban-like settlement.
Already at this early time, the third term, nagara, might have had a wider connotation, referring to the city and the hinterland controlled by it. The question as to whether the epigraphical evidence allows such a terminological distinction between the kraton, the town, and its politically controlled hinterland will remain a crucial problem throughout this article. One point, however, is worth mentioning in connection with the nagara of Pΰrnavarman‘s inscription. The fact that only the term nagara occurs in connection with Taruma and the polities of its defeated enemies may be understood as an indication that the spatial concept of the "state" in fifth-century Java was primarily "city"-centered. The earliest epigraphical evidence of Indonesia from the fifth century thus confirms Wheatley‘s definition of early political units on the Malay peninsula which he derived mainly from Chinese sources: "a polity in which a focally situated settlement exercised direct control over a restricted peripheral territory and exacted whatever tribute it could from an indefinite region beyond." [13]
II
The next evidence of early urbanism and state formation in Indonesia comes from the most interesting corpus of inscriptions of the early Malay world, i.e. the inscriptions of early Srivijaya which can be dated around 682 to 686 C.E. The spatial distribution of the seven major inscriptions indicates an interesting pattern. Three have been found at Sri vijaya‘s center around Palembang [14] and four were discovered in outer regions {mandala) [15] that encircled the center at a distance of several hundered kilometers. Furthermore, some fragmentary inscriptions have come to light in the center. [16] This distribution pattern foreshadows a major problem of Srivijaya‘s statehood, the control of its outer regions. The most important inscription is the Telaga Batu or Sabokingking (Skk) inscription of Eastern Palembang, which contains dreadful curses against disloyal members of the royal family and local chiefs. The author of all these inscriptions seems to have been king Jaynasa under whom Srivijaya experienced a period of breath-taking expansionism, conquering Malayu-Jambi and Kedah and attacking Java with a naval expedition. [17]
The Skk inscription begins with "an almost entirely unintelligible oath formula” [18] which is followed by a peculiar list of officers and occupational groups. [19] It includes princes (jάjaputra), landlords (bhΰpatϊ), army leaders (senάpati), local magnates (nάyaka), confidants (φratyaya), royal confidants (haji pratyaya), [20] judges (dandanάyaka), surveyors of groups of workmen (tuhά an vatak = vuruh), surveyors of low castes (addhyάksi nϊjavarna), cutlers (vάsϊkarana), ministers of princely status (kumάrάmάtya), regular and irregular troops (cάtabhata), administrators (adhikarana), clerks (kάyastha), architects (?, sthάpaka), naval captains (puhάvam), merchants (vaniyάga), royal washermen (marsϊhάjϊ), and royal slaves (hulun haji). De Casparis is certainly right to assume that this heterogeneous list contains those "categories of people that might constitute a possible danger" [21] to the security of the king and his court and who had therefore to swear the oath. The inscription furthermore refers to three categories of princes: the crown prince (yuvarάja), second crown prince (prάtiyuvarάja), and other princes (rάjakumάra).
These lists seemingly depict a well-organized hierarchy of princes, court officers, and servants at Sri vijaya. As no identical lists are known from contemporary India or Southeast Asia it is likely that they reflect a fairly true picture of an already quite advanced society at the court of Sri vijaya. However, it may have been differentiated more horizontally than hierarchically structured. This assumption is based on the fact that the many Sanskrit titles of court officers are known only from the introductory list of the Skk inscription, which, later on, also thrice mentions the three categories of princes. Otherwise two other Malay terms were apparently considered much more important than these "foreign" names. These were dάtu and huluntuhan. Dάtu appears to be the traditional Malay title of a chief and occurs frequently in all Malay inscriptions of early Sri vijaya. In the mandala inscriptions even the "King" of Sri vijaya is referred to by this traditional title. [22] Huluntuhan, the "slaves [hulun] and lords [tuhan]," occur seven times in the Skk inscription. This term obviously denotes the members of the traditional patriarchal household of the dάtu of Sri vijaya and therefore would have included all sorts of family members and retainers who acted on behalf of the dάtu of Sri vijaya. These huluntuhan most likely were identical with those officers mentioned in the introductory list.
As regards the spatial dimension of Srivijaya‘s political authority these Malay inscriptions contain several new pieces of important information that can be derived from several key-words in these inscriptions, in particular kadάtuan, vanua, samaryyάda, mandala, anbhϋmi. In a recent paper [23] 1 have tried to show that these terms allow us to draw some conclusions on the early process of concentric state formation in seventh-century Sumatra. The center of this process was the kadatuan Srivijaya, "the place of the datu" of Srivijaya, which had thus the same meaning as Javanese keraton, "the place of the ratu." [24] According to the Skk inscription the kadatuan contained the inner apartments of the residence (tnah rumah) of the datu where his women folk (bint hajί) lived and where gold (mas) and tribute (draxvya) were kept. Most likely the kadatuan also housed the dewata, or tutelary deity (a deified ancestor?), which, according to the mandala inscriptions, protected the kadatuan of Srivijaya.
The kadatuan was surrounded by the vanua Srivijaya, the semi-urban area of Srivijaya. [25] Fragments of inscriptions refer to citizens (paura) and to a "vihara in this vanua” [26] This monastery may have housed some of the one thousand Buddhist monks whom the Chinese monk I-ching mentioned as being in Srivijaya during these years. Furthermore, the Buddhist park Sriksetra established by King Jayanasa and Bukit Senguntang, Siϊ vijaya‘s "sacred center" [27] where a number of Buddhist remains have been unearthed, may have been situated within this vanua Srϊvίjaya. Morever, we may assume that it contained several truly agricultural villages, and, in particular, the markets frequented by the foreign merchants (vaniyaga) and sailors (puhavam) who are mentioned in the Skk inscription. The kadatuan and the vanua formed the nucleus of Srivijaya. Only these two terms occur in connection with the name Srivijaya [28] and, what is perhaps equally revealing, only these two central spheres are known by their Malay terms. The other more distant surrounding "circles" of this nuclear area are referred to in the inscriptions only by the Sanskrit names samaryyada, mandala, and bhΰmi.
The term samaryyada is very unusual; in fact, it is unknown in Indian [29] or Southeast Asian epigraphy. De Casparis translates the term as "frontier province" whereas I prefer its literal meaning "having the same boundaries" (maryyάda). The samaryyada thus would have referred to the neighboring region beyond the vanua Srivijaya. According to the Skk inscription the samaryyada was connected with the central vanua by special roads (samaryyάdapatha). This hinterland was populated by an obviously large number of datu who resided— according to the Skk inscription—in their own places (sthάna) on their own land (desa). But these local dάtu have come under the control of the datu of Srivijaya and his huluntuhan who appear to have been particularly active in the samaryyada hinterland. The endeavor of these datu to again become "independent" (swasthd) must have posed one of the greatest dangers to the security of the datu of Srivijaya and the kadatuan as these datu are threatened several times by the imprecations of the Skk inscription.
The fourth important term for our consideration about the spatial dimension of Srivijaya‘s statehood is mandala. It occurs only once in the Skk inscription in the famous passage mamraksaήa sakalamandala kadatuanku, which de Casparis translates "you who protect all the provinces of my empire” [30] 1 have tried to show that, in the context of early Sri vijaya, the term mandala most likely did not refer to centrally administered provinces in the extended core area of an empire, as such provinces did not yet exist in early kingdoms; they became a typical feature only of the later imperial kingdoms. [31] In the context of early kingdoms, the term mandala usually referred to autonomous or semi-autonomous principalities and chiefdoms at their periphery. Several such mandala are known from contemporary Southern and Eastern India. [32] In the case of late seventh-century Sri vijaya "all the mandates" therefore appears to have meant those outlying regions where the abovementioned "mandala inscriptions" have been found. The exact nature of Srivijaya‘s control over its mandala is unknown. They had certainly been conquered by Srivijaya‘s army (bala), which is mentioned several times in Srivijaya‘s inscriptions. However we have no evidence at all that they had come under the direct political control of Siϊvijaya as no huluntuhάn or royal princes of Srϊ vijaya are mentioned in these mandala inscriptions. They were obviously still ruled by the local dotu who lived in their own vanua, as is known from these inscriptions. They had been recognized (sanyάsa) in their position by Sή vijaya. But their precarious loyalty obviously had to be improved by the mandala inscriptions and their peculiar mixture of imprecations and taking of the oath of allegiance. However, military coercion and imprecations alone would not have sufficed to establish an enduring relation. Of equal importance must have been the incentive to participate in Srivijaya‘s international trade.
Apart from providing us with a conceivable model of an early concentric state, SIΪ vijaya also provides us with the first generic term of such a state. As mentioned above, Coedes, de Casparis, and Boechari regarded vanua, kadάtuan, and huluntuhάn as just such a comprehensive term and translated them accordingly as "le pays," "empire," or "kingdom." But, according to my interpretation, none of these expressions had such a comprehensive spatial connotation in the context of early Sri vijaya. The word bhΰmi, however, appears to havebeen such a generic term for Srivijaya‘s statehood. In Sanskrit, bhΰmi means primarily "earth" or "soil" but also "realm" and "country." It occurs twice in Srivijaya‘s inscription. One instance is in a more or less identical passage found in all the mandala inscriptions, which threatened the disloyal "people inside the land [that is] under the order of kadάtuan"(urah dί dalahήa bhΰmi άjήάήa kadatuanku). [33] As this passage occurs only in the mandala inscriptions it has to be inferred that the places where they have been found either constituted a bhΰmi or formed part of a larger polity which was called bhΰmi. Although the first meaning cannot be excluded, two other references make the latter connotation of bhΰmi morelikely in the context of early Indonesian history. The first of these references comes from the important passage of the Kota Kapur inscription of the year 686 which announces the departure of an army expedition against bhΰmi jawa, which had not yet become submissive to Sri vijaya. Obviously, Java was not regarded just as one of the many vanua or mandala surrounding Sri vijaya but as an equally matched opponent of Sri vijaya. The other evidence of a bhΰmi polity comes from several inscriptions of late ninth and early tenth-century Java that refer to bhΰmi Mataram (infra). As in the case of bhΰmi Java and bhΰmi Mataram, the Srivijayan concept of "the bhΰmi under the control of my kadatuan" apparently referred to the whole sphere that had come under the control of Sri vijaya.
Comparing Mΰlavarman‘s and Pΰrnavarman‘s fifth-century polities with Srϊvijaya in the late seventh century we are able to recognize several important structural changes. The "first victories" (prathama vijiya) of these earliest kings certainly had led to the defeat of neighboring "landlords" (pάrthίva) and "hostile towns" (arinagara). And in some cases a victorious "lord of the landlords" (pΰrthίva-indra) may have been able to collect (most likely irregular!) tribute (kara). But none of their earliest inscriptions allow us to infer that this pristine political development presupposed or led to far-reaching structural changes. The courts of these early rulers still remained patriarchal households. Their rule was the affair of the chiefs family (kula) or lineage/"dynasty" {vamsa). It is this background that explains the frequent mention of these two kinship terms in these early inscriptions, that is, kula and, particularly, vamsa. However, it is worth mentioning that these Sanskrit terms only occur whenever the foundation of a "dynasty" is reported in inscriptions. As no equivalent Malay or Javanese word ever occurs in inscriptions written in these languages, it is likely that these Sanskrit terms were required for the definition of an apparently new social institution, that is, a "ruling lineage" in a hitherto rather unstratified tribal society.
This situation had already changed considerably in late seventh-century Sumatra around present-day Palembang. Whether international trade by the "sailors and traders" (puhavam vaniyaga) mentioned in the Skk inscription was the main cause of this change is still an open question. But it is evident that the inscriptions of Sri vijaya which have been discovered around Palembang depict an already fairly well-developed society. But in this regard, too, we have to distinguish between different spatial zones of social change. Social differentiation and stratification was strongest in the vanua, center from where its influence spread into its samaryyada hinterland. But we have no evidence of such a development in the outer mandala, even though we may conjecture that the dάtu of these regions had their own huluntuhάn. But particularly in these cases they would have been patriarchal "servants and lords" rather than administrative officers.
Apart from new social stratification in vanua Sri vijaya, the other decisive difference between the epigraphical evidence of fifth-century Kalimantan and Java on the one side and late seventh century Sri vijaya on the other side is an incipient change in the center-periphery relations. Sri vijaya seems to have been the first Indonesian state that succeeded inextending its direct political authority beyond its own vanua into the samaryyada hinterland and to conquer even far-off powerful chieftaincies and trade emporia (e.g. Malayu and Kedah) and to establish some sort of hegemony over these outer mandala. Srivijaya‘s rapid expansion was due to two new factors which had different spatial significances. First, Srivijaya‘s direct rule in the samaryyada hinterland was based primarily on its disposal of a fairly well-developed staff of "administrators," the huluntuhan of the Skk inscription. However, these "officers" still had to face various types of difficulties at the center and in its hinterland that are vividly accursed in the Skk inscription. As shown by J.G. de Casparis in his analysis of the fragmentary inscriptions, vanua Sri vijaya was still afflicted by dangerous rebellions of internal insurgents. [34] Second, Srivijaya‘s control over the far-off mandala presupposed the existence or at least the temporary availability of a strong army (vala). Its inscriptions show that the main cause of Srivijaya‘s hold over the outer mandala was the ability to muster an army of apparently uncontested strength. Nevertheless, Srivijaya‘s authority still remained very precarious in these mandala. Srivijaya‘s major problem at this point appears to have been its "failure" either to integrate at least some of the more powerful datu chiefs of the mandala into its own central court or to obliterate and replace them with loyal members of its own court. In fact, the solution of these center-periphery relations remained the crucial problem of all pre-modern states. As we will see, the Mataram-Sailendra dynasties chose the first and Majapahit the second method to solve this problem.
In regard to the main concern of this article, the epigraphical evidence for the "city" and the "state" in early Indonesia, Srivijaya provides for the first time a rather clear picture of an already well-developed concentric state in early Indonesia. Mύlavarman‘s polity had comprised only two such spatial spheres. These were a still rather undifferentiated center consisting of a pura (= kraton?), a "most holy place" (ksetra), and the living places of the "royal" vamsa and the Brahmins. Beyond this center a group of unspecified defeated chiefs (pάrthiva) existed whom Mϋlavarman claims to have made his "tribute-givers" (karadά). Pΰrnavarman‘s inscriptions from West Java offer a slightly more-developed stage of an early polity. As all three epigraphical terms that occur in these inscriptions and that are relevant for our study, that is, sibira, purϊ, and nagara, have a clear urban connotation, Tarumanagara appears tohave been a typical early "city state." But we have to keep in mind that in the context ofearly Indonesia these Sanskrit terms may have referred to kratons of strong chiefdoms and Hinduized "little kings" rather than to an urban "city."
The major contribution of Srivijaya‘s inscriptions to our study is the fact that they indicate several spatial spheres of its political authority. And moreover, for the first time in early Indonesia, a generic term, bhΰrni, for the "state" is given. This bhΰmi realm was divided into three zones, that is, the vanua center, its samaryyάda hinterland, and the outer mandala. The center, too, consisted of two distinct zones, the kadάtuan (or kraton) of the ruling datu and its surrounding vanua. Vanua Srivijaya seems to have comprised a densely populated area at present-day Palembang with urbanized "pockets." However, it is astonishing that we only once hear of Srivijaya‘s "citizens" (paura) whereas the terms pura or nagara never occur in Srivijaya‘s inscriptions. Whether this evidence is sufficient to infer that no "city" outside the kadάtuan Srivijaya existed in the late seventh century is difficult to decide. But the epigraphical evidence makes such an inference quite likely. [35] A possible reason why we may search in vain for the epigraphical terms pura and nagara in the context of early Srivijaya may be the fact that much of its urban life took place in the many houseboats on the Musi river.
III
The two earliest dated inscriptions of the middle period of early Indonesia, the Canggal nscription of Sanjaya of the year 732 C.E. [36] and the Dinaya inscription of 760 C.E., [37] belong to Central and Eastern Java, respectively. Their date is only about two generations later than Sri vijaya‘s early inscriptions. But in regard to Java itself, there is a wide gap of three centuries between Pΰrnavarman and Sanjaya. Both inscriptions of the eighth century provide some new evidence of a conceptual and structural development of the "city" and the "state" in early Indonesia. Sanjaya‘s inscription reports the consecration of "a liήga on a mountain" (prάtisthipat parvate lihgam) in a place or "country" (desa) called Kunjarakunja in the island (dvtpa) of Java. Sanjaya is praised for protecting the "royal highways" (rάjapathi), defeating "numerous circles of neighboring chiefs" (aneka-samanta-cakra-mja), and for ruling justly his kingdom (rajya). From later inscriptions we know that the Mataram dynasty praised him as "king" (ratu) and founder of their dynasty. [38] Furthermore it is worth mentioning that Saή- jaya‘s Canggal inscription provides the first epigraphical evidence for agrarian extension when it mentions that Java was "rich in rice and other seeds."
The evidence of Sanjaya‘s inscriptions as well as that of later inscriptions leaves no doubt that Sanjaya established a genuine early kingdom, in fact the first known in Javanese history. It is of particular interest for our study that the concept of his state and the prasasti eulogy of the inscription were much nearer to contemporary Indian models than the earlier cases that so far have been discussed. For the first time, the "state" is called rάjya. It comprised, most likely, several desa which were linked by "royal highways." It was surrounded by a sάmantacakra, the common term of Indian and future Javanese inscriptions referring to the "circle of (originally independent) neighboring rulers (sάmanta)." [39] According to the Indian concept they had to be subjugated to the central king. Another indicator of Indian influence is the consecration of a purely Hindu temple, dedicated to Siva.
But despite this evidently strong Indian influence, Sanjaya‘s inscription lacks several important indicators of contemporary Indian statehood and kingship. This is most evident in regard to the complete absence of officers usually mentioned in connection with such a grand ceremony as the establishment of Sanjaya‘s "royal" Sivalmgam. Whether this evidence allows the argumentum ex silentio that Sanjaya‘s court still consisted only of his extended family (kula), which in fact is mentioned twice in the inscription, is a matter of conjecture. But in this regard it has to be remembered that Sanjaya‘s "dynasty" was not yet an old-established one. He traced his genealogy back only to his father Sanna, and he himself is praised in a later inscription by the Malay title ratu whereas his successors are glorified as mahΰmja. [40] In this regard it is significant that his father is reported to have ruled his people by the traditional method of "conciliation and gifts" (sama-dana). Another interesting point is the fact that Sanjaya‘s inscription does not contain any term that can be associated with an urbanized settlement. This again does not automatically mean that no such settlement existed in his rajya in Central Java. But at least this evidence shows that he did not regard it necessary (or important enough) to mention such a settlement in his inscription.
A rather different picture of an early kingdom is depicted in the Dinaya inscription of the year 760 C.E., the earliest one known from East Java. Much has been written about it because of several uncertainties about its contents, which, however, need not bother us here. [41] It reports the consecration of a temple and a stone image of the divine seer Agastya by king (raja) Gajayana who protected a pura, called Kanjuruha. King Gajayana urged his relatives (bandhava), sons (nrpasuta), and his principal ministers (mantri mukhya) not to act against this gift. He appealed to the members of the "royal dynasty" (vamsa nrpa) to follow his example and to perform meritorious acts and thus to protect the kingdom (rάjya).
The interesting evidence in regard to the "city" of Kanjuruha is the fact that the first verse mentions that the pun was protected by Gajayana‘s deceased father Devasimha, whereas the next verse reports that Gajayana protected the pura Kanjuruha after his father had died. Furthermore we are told about "citizens" (paura) who, together with the "groups of leaders" (nayaka-gana), constructed the temple. The juxtaposition of pun and pura in the first two verses, which were protected by the deceased father and his ruling son respectively, allows us to infer that in this case pun may have referred to the kraton and pura to the "city" of Kanjuruhan. The pun thus appears to have been protected by the (deified) ancestor (as was kadatuan Srivijaya by its devata), whereas King Gajayana protected the pura "capital" of his rajya. It was inhabited by paura, groups of nayaka, ministers (mantri), and Brahmins who, too, are mentioned in the inscription.
In contrast to Saήjaya‘s inscription from Central Java, the Dinaya inscription from East Java thus reveals an urban center comprising the pun -kraton and the pura. The latter was inhabited by members of the royal family, by "citizens," officers, and priests. However, although Gajayana‘s "state" is called a "kingdom" (rajya), the inscription does not contain a single piece of evidence that would allow us to assume the existence of administrative units. This fact reminds us of Sanjaya‘s nearly contemporary rajya where we came to know about the existence of at least one desa. Despite the mention of "principal ministers" and "groups of leaders," and "citizens," the strong emphasis of the responsibility of "royal relatives," princes, and members of the "royal dynasty" for the welfare of the rάjya makes it likely that Gajayana‘s "state," too, consisted mainly of an urbanized center that may have had strong relations with, but only little or even no political control over, its surrounding hinterland. The major difference between Mϋlavarman‘s polity in the fifth century and Gajayana‘s eighth-century "kingdom" thus appears to pertain to the degree of urbanization of its center rather than to the development of its territorial dimension.
IV
The late eighth century bears witness to the rise of the Buddhist Sailendras, one of Southeast Asia‘s most important dynasties to which the world owes one of its greatest religious monuments, the Borobudur. But despite its importance in early Indonesian history and its historically established links with Sumatra, the Malay Peninsula, Sri Lanka, Bengal, and South India till the eleventh century, its genealogical history and, even more, the structure and extent of its kingdom, particularly during its most important period in Central Java in the eighth and ninth centuries, is only partly known. [42]
The major reason for our lack of knowledge is, no doubt, the relatively small number of inscriptions. But in the same way as the international relations of the Sailendras and their masterpieces of art and architecture indicate a new stage of cultural, societal, and political development, the dynasty‘s few inscriptions, too, reflect a new type of full-fledged kingdom. Already the earliest-dated Sanskrit inscriptions of the eighth century speak of a rajya state, ruled by a maharaja who had defeated the neighboring samanta-mja. Ministers (mantri), (local?) lords (pati), and superintendents of desa areas (desάdhyak$a) were in charge of the administration.
In this connection it is of particular importance that the Kalasan Sanskrit inscription of the year 778 C.E. mentions for the first time the local desa officers paήgkura, tavάn, and tϊrip which belong to the large number of different Javanese titles of officers that occur frequently in Javanese inscriptions of the ninth century and later. The obviously well-established hierarchy of officers, linking the central court with intermediary, territorially defined, administrative units and villages, is certainly the most important new epigraphical evidence of the classical age of central Javanese history under the Sailendra and Mataram dynasties. This is not the place to discuss in detail this complicated administrative set-up of Central Javanese kingdoms. [43] Suffice it to say that it appears to have been the outcome of a double process that must have been operating for a much longer time than evidenced in the inscriptions. On the local level of the vanua, agrarian expansion, translocal trade, and social differentiation had created a vast number of village elders and authorities. The early history of this process is unknown to us as we come across its result only when these vanua authorities were already fully existing in Central Javanese inscriptions in the early ninth century. The other aspect of this process is of equal relevance for our study as it pertains to the political expansion of supra-local authority. Mΰlavarman‘s and Siΐ Mara‘s inscriptions illustrated the incipient stage of this process. The inscriptions of Pΰrnavarman, Sanjaya, and Gajayana depicted further stages of this development. But in their cases, too, no political authority appears to have as yet been established permanently beyond the chiefly "Stammland" and its immediate hinterland. This decisive step of early state formation was made by Srivijaya which extended its political control into distant mandala. However, Siϊvijaya‘s authority in these mandala continued to be precarious, as their local datu leaders remained a threat to Srivijaya‘s datu rather than being integrated into the political structure of the bhQmi state.
According to our epigraphical evidence, political expansion through integration occurred only with the rise of the Sailendra dynasty and reached its first culmination towards the end of the central Javanese period in the late ninth century. [44] It is worth noticing that in contrast to the Sanskrit inscriptions of the early period of Java, this ninth-century process of intensive state formation in Central and, from the tenth century onwards, East Java is documented nearly exclusively by Javanese-language inscriptions with but little Sanskrit terminology. The basis of this expansion of royal authority was the stepwise integration of neighboring areas. Even though they remained "under the jurisdiction" (watek) of local chiefs (raka), some of these chiefs slowly rose to high administrative or even "ministerial" positions in the patriarchal central court. In the same way as these chiefs retained their "Stammland" as their own watek, the king (maharaja, ratu) kept his own watek land under his direct control. Although having thus, at least theoretically, the same territorial basis of authority, the "Stammland" of the future ratu may have been larger, perhaps more fertile, and linked to translocal trade routes. Certain "material" factors must have given the family/ lineage of the future ratu an advantage over his neighboring chiefs (raka) during this process of early state formation in Central Java.
As already mentioned, the early history of this process of territorial and political integration of the raka into a central court of a ratu is largely unknown, as these raka are usually referred to only in the inscriptions of a central ratu when a certain degree of their integration had already occurred. However, among the early inscriptions of classical Central Java several inscriptions are known to have been edited by local chiefs without referring to a ratu or raja. A rare case even allows some conclusions about the process of political integration when the Central Javanese court was just moving to the east. In 891 C.E. the rakryan of Kanuruhan (a locality most likely identical with Kanjuruhan of the Dinaya inscription of 760 C.E.) established a freehold by his own "favor" (anugraha) [45] But only twenty-four years later, in 915 C.E., it was the central "Great King" (mahάmja) Daksa who did the favor (anugraha) to allow the rakai of Kanuruhan to establish another freehold in his own watek Kanuruhan. [46] Between 891 and 915, under circumstances still unknown, the Lord of Kanuruhan thus had come under the authority of the Maharaja Daksa. [47] During the next century the rakryan of Kanuruhan rose to the highest administrative position at the central court, a position which they held for several centuries.
The stepwise integration of local magnates into the central court and the encroachment of the royal "persons who collect the lord‘s property" (mangilala drawya haji) upon the watek and vanua is perhaps the least known aspect of state formation in early classical Javanese history. But it seems to have been an extremely protracted process that finally worked in favor of the central dynasty only temporarily in the fourteenth century, when Majapahit was able to exchange the local raka in the extended core area of its kingdom with members of its own dynasty.
The structure of the Javanese kingdoms of the ninth through early thirteenth centuries is fairly well reflected in the epigraphical evidence and may be summarized as follows. The unique feature of the Javanese kingdom during this period is the highly elaborated and strictly fixed hierarchy of state officers and local authorities as listed in these inscriptions. The establishment and continuous maintainance of this hierarchy appears to have been one of the main means of authority for the center. This hierarchy of patrimonial officers, however, should not be automatically equated with a hierarchy of administrative officers and administrative territorial units. Apart from "personal" watek and from vanua with their village authorities, which are frequently mentioned in inscriptions, we still find nearly no terms that hint at a more sophisticated spatial structure of the kingdom. Very rarely does the term desa occur, which refers to a territorial unit of a larger size than vanua. An inscription of the year 824 C.E. appears to indicate that the rάjya constituted of many desa. [48] The Sanskrit term desa may thus have to be equated with watek. [49] This inference perhaps may be corroborated by another early Central Javanese inscription of the year 782 C.E. It reports the consecration of a Manjusrϊ image by a Sailendra king and his guru who had come from Bengal (gaudidvtpa). The image was installed by the king in order to protect "his desa" (desasya tasya). [50] Here desa apparently refers to the land of the king, that is, his Stammland or watek.
Important information about the structural concept of the Javanese state during this period can be derived from a passage which is repeated more or less identically in a few inscriptions of the late ninth and early tenth centuries. It occurs for the first time in an inscription of King Lokapala of the year 880 C.E. [51] At the end of an extremely long list of Hindu deities who are called upon for protection of a newly established freehold, the tutelary deities are invoked: "Also all you deities who are known to protect the kraton of the illustrious Great King in the country of Mataram" (devata prasίddha mangraksa kadάtuan srί maharaja i bhΰmi i mataram). Nearly exactly the same text is repeated in the same epigraphical context in the stone inscription of Sugih Manek near Singasari of the year 915 C.E [52] Thirteen years later this invocation occurs again in the so-called Minto Stone from the region of Malang, however with an important addition. [53] The tutelary deities are invoked "who are known to protect the kraton of the Great King in Medang in the country of Mataram” For the first time, Medang, the capital of the central Javanese kingdom of Mataram, is explicitly mentioned in this context. This latter version of the invocation is then again repeated twice in inscriptions of the years 942 and 944 C.E. [54]
In the context of our study, the most relevant aspect of these inscriptions is the fact that they reveal, at least partly, a repetition of the Srivijaya model. As in Siϊvijaya, the center was the kraton (or kadatuan) of the Maharaja, which was protected by ancestor deities. It was situated in (the "capital") Madang which may have been identical with vanua Mataram, a resident of which is mentioned in an inscription of the year 919 C.E. (anak vanua i mataram). [55] This center was surrounded, as we know from other contemporary inscriptions, by the desa or watek of other raka or rakryan. The state that comprised these "segments" was called bhΰmi, a term we came to know for the first time in Siϊvijaya‘s inscriptions. The most important and, in contrast to early Srivijaya, new political element of state formation in pre- Majapahit Java is the obviously very successful integration of allodial chiefs and "lords" into the patrimonial hierarchy of the central court without, however, uprooting them in their own Stammland [56] No such attempts to integrate the datu into the court hierarchy are known from the inscriptions of Srivijaya. In contrast to bhΰmi Srivijaya, however, bhΰmi Mataram appears not to have included outlying mandala.
Furthermore, the territorial administration of bhΰmi Mataram in Central Java and its successor kingdoms in East Java may have been structured even less than Srivijaya in the late seventh century. At least, we have no inscriptional evidence, for instance, of the kingdom (bhΰmi) of Kediri from the late eleventh to early thirteenth centuries which would allow us to come to a different estimation of its statehood. There seem to have been only two exceptions to this "rule," that is, king Sindok in the early tenth century and Airlangga in the early eleventh century. They obviously had tried to extend their political control even beyond those watek or desa that had already come under their hegemony. But their personal, most likely military, success did not survive their demise. Airlangga‘s inability to perpetuate his temporary success was later on transformed into the famous myth of the division of the "empire" by the king himself. However, Airlangga‘s striving for "imperial" hegemony (he was the first Javanese king who assumed the imperial title ratu cakravartin [57] ) became a major prop of the imperial ideology of later Singasari and Majapahit rulers.
During the Mataram and Kediri period, the "bhΰmi state" of East Java thus remained the classical concentric "Early Kingdom." [58] It comprised the central nuclear area, consisting of the kraton and the Stammland or watek of the central dynasty, and the surrounding watek whose rakryan and rakai had been integrated into the patrimonial staff of the central court. The local administration in the royal and allodial watek remained in the hands of the traditional village authorities.
Finally it should be pointed out that throughout the period of the Sailendra dynasty and the kingdom of Mataram and Kediri, no epigraphical evidence exists about the existence of "cities" except those of the "capitals" of these kingdoms, e.g. Madang and Kediri. But neither are they called pura or nagara nor are urban settlements known in the watek of the rakryan. According to the epigraphical evidence, state formation in Java in the eighth to early thirteenth centuries appears to have operated largely without urbanization.
V
The most dramatic changes in the process of state formation and urbanization in pre- Islamic Indonesia took place in the late thirteenth and fourteenth centuries under Singasari and Majapahit. For nearly two centuries after Airlangga, the history of Java had retreated to the middle Brantas valley. However, the small but obviously prospering kingdom of Kediri preserved faithfully the structural concept of the administrative set-up of the Mataram kingdom and the imperial ideology of Airlangga‘s short-lived East Javanese kingdom. This ideology was based on the (under Airlangga only partly achieved [59] ) unification of the two East Javanese nuclear areas Janggala and Panjalu and the (never-realized) hegemony over the whole of Javadvϊpa. The myth of a unified kingdom of Janggala and Panjalu (and its alleged partition by Airlangga) and East Java‘s imperial claim over the whole of Java served as legitimation of the imperial expansionism of late Singasari which Coedes summarized as follows:
The reign of Krtanagara [the last king of Singasari] was marked by a considerable expansion of Javanese power in all directions. In 1275, taking advantage of the decline of Srivijaya, he sent a military expedition to the west which established Javanese suzerainty over Malayu and probably also over Sunda, Madura and part of the Malay peninsula. After establishing his authority in Sumatra, Krtanagara turned toward Bali, whose king he brought back as a prisoner in 1284 [60]
In East Java itself, the most important change was the subjugation of Kediri under Singasari‘s hegemony, which had already been finally established under Visnuvardhana, Krtanagara‘s father. However, Kediri still retained an autonomous status as a samantamjya [61] King Jayakatwang of Kediri still felt strong enough to attack and defeat Singasari in 1292 C.E. During the occupation of the royal residence of Singasari, King Krtanagara died. Kediri‘s new hegemony over the whole of eastern Java however came to an abrupt end when Krtanagara‘s son-in-law, Raden Vijaya, with the help of a Chinese expeditionary army, defeated Jayakatwang‘s army and established himself as the first ruler in the newly founded city of Majapahit.
The kingdom of Majapahit was truly a successor state of Singasari, both in its "internal" policy in eastern Java and in its expansionistic "external" policy in the outer regions. But in both aspects the state of Majapahit represents the culmination of state formation in premodern Indonesia. As regards the "internal" policy in eastern Java, the most decisive new development under Majapahit was the systematic replacement of the allodial local raka and rakryan in East Java by members of the royal family and, in a few cases, by deserving members of the court. It was more than a mere symbolic act that, already in 1295, Raden Vijaya crowned the eldest son of one of his wives (he was married to four daughters of Krtanagara) as Prince of Kediri. His new policy of systematic annexation or "provincialization" of all neighboring watek or desa and their many "little kings" led to a series of revolts, which, however, appear to have been successfully supressed by Majapahit. According to our epigraphical evidence, Majapahit succeeded for the first time (after Srivijaya‘s similar attempts) to extend its political control considerably beyond its own Stammland, this time, however, by a ruthless policy of annexation and "dynastification" of its hinterland.
In this regard it is significant that the Nagara-Kertagama, the famous court chronicle of Majapahit composed by Prapanca in the year 1365 C.E., begins with a descriptive list of the various towns (nagara) held by members of the royal family as demesne in the hinterland of Majapahit. Summarizing this chapter, the Nagara-Kertagama concludes "All Illustrious Javanese Kings and Queens, the honoured ones who equally are distinguished by their towns (nagara), each having one for his or her own, in one place (eka sthana), in Wilwa Tikta (= Majaphit) they hold in their lap the honoured Prince-Overlord." [62]
As is known from other Southeast and South Asian kingdoms of this age, too, (e.g. Angkor and the Colas), the establishment of an "Imperial Kingdom" required a considerable enlargement of the original nuclear area or Stammland of the ruling dynasty. [63] In this extended core region the ruling dynasty had to strive for uncontested access to the agrarian surplus and, wherever possible, for some sort of control over, and sharing of, the long-distance trade. However, even in these "Imperial Kingdoms" of pre-modern South and Southeast Asia, the central dynasties still had to share the revenue from these sources with their own local representatives, whether they were princely members of the dynasty or members of the patrimonial staff. The transfer of resources from the local and intermediate levels to the imperial center thus remained a crucial problem even within the core region of these large imperial kingdoms. Although the imperial kings had succeeded in extending their uncontested political authority by eliminating all sorts of potential putschists in their extended core region, actual political control remained fragmented. The "segments" still existed even though they had come under members of the central dynasty. [64] But in contrast to the earlier cases (e.g. in 760 C.E. in Kanjuruhan in East Java) where a "family dynasty" (kula-vamsa) depended mainly on its own Stammland, the imperial dynasty of Majaphit was able to extend its patrimonial control far beyond these pristine boundaries. It systematically distributed its hinterland to members of the dynasty who thus became "share holders" of the state. The central core region of Majapahit had come under a "family regime."
The very center of this family enterprise was the royal compound (pura) or kraton. Accordingly the second chapter of the Nagara-Kertagama contains a detailed description of this compound and the surrounding nagara, that is, the capital of Majapahit. The Nagara- Kertagama again pays special attention to the residences of all the principal ministers and the princely family members whom we have already met in their own nagara. According to their status, their residences were distributed in a clear hierarchical order around the royal pura or kraton, adjacent to temples, monasteries, markets, and places of the commoners. The Nawanatya, a most-likely much later text, contains a nice definition of the nagara. "What is called nagara? All where one can go out (of his house) without passing through paddy fields” [65] Archaeological surveys at present-day Trowulan and literary evidence confirm that Majapahit was a truly urban settlement, in fact the earliest in Java so far known both from archaeological and literary sources.
In regard to the spatial concept of statehood the Nagara-Kertagama and an inscription of 1323 C.E. from Tuhanaru contains an interesting piece of information. [66] In this inscription the kingdom (rάjya) of Majapahit is compared with a temple (prasάda) in which the king is worshipped as an incarnation (avatάra) of Visnu. The mandala of the island (dvϊpa) of Jawa is equated with the temple land (punpunan) whereas the islands (nϋsa) of Madhura, Tanjunpura, etc. are compared with amsa land or dependencies which were only partly (amsa) under the control of Majapahit. The interesting point is the fact that this inscription clearly distinguishes between the rάjya Majapahit and the (surrounding) mandala of Jawadvϊpa. It therefore appears that only the core region, comprising the kraton, the capital, and the nagara of the princes, constituted the rajya of Majapahit. We are used to translating this term as "kingship" or "kingdom" as in the literal sense it means "belonging to the king." Therefore it should be no surprise that in the context of South and Southeast Asian concentric states, rajya actually referred only to the inner core region under the direct authority of the raja.
Beyond this rajya of Majapahit in eastern Java was bhΰmί (or dvϊpa) jawa, which, for the first time in its history, has come under the hegemony of a single dynasty. No sources, however, are available that would allow us to infer that here in Jawadvϊpa-Mandala, too, members of the central dynasty or court have been imposed "from above" as local rulers. Outside the rάjya of Majapahit, bhΰmί jawa apparently was still under the watek of its autonomous local raka. Moreover we have no idea as to whether Majapahit was able to establish any sort of provincial administration in these autonomous watek of bhΰmi jawa outside the rάjya of Majapahit and to collect regular taxes in these regions. But militarily, Java was certainly fully under the control of Majapahit. Furthermore, we may assume that military expeditions and visiting officers of the central court had to be supplied by local authorities.
Beyond bhΰmijawά were the "other islands" (nusάntara or dvϊpάntara). The Nagara- Kertagama, 13-16, contains a long list of these islands, which include most of present-day Indonesia‘s islands as well as parts of the Malay peninsula. Most important among these islands was Sumatra with Malayu, the successor state of Palembang/Srϊvijaya. It was the only polity on these outer islands to which the Nagara-Kertagama concedes the important term bhΰmi. This reminds us of the fact that in Siΐvijaya‘s early inscriptions, too, the term bhΰmi was reserved for Srϊvijaya and Java. The outer islands were regarded as tributary states of Majapahit. The Nagara-Kertagama claims that "already the other islands (dvTpάntara) are getting ready to show obedience to the Illustrious Prince, without exception they bring in order all kinds of products every ordained season. As an instance of the honoured Prabhu‘s [King‘s] exertion for all the good that is in his care, ecclesiastical officers (bhujahga) and mandarins (mantri) are sent to fetch the produce regularly” [67]
The last concentric circle of Majapahit‘s statehood was constituted by the "other countries" (desάntara). According to the Nagara-Kertagama, Siam, Ayuthaya, Ligor, Martaban, Rajburi, Singhanagarϊ (=Satingpra), Campa, and Kamboja belonged to this category; Yavana (= Vietnam) "is different, it is a friend (mίtra)." [68] The desΰntara countries most likely were identical with "all the mandalita rastra (which are) looking for support, numerous, entering into the Presence." [69] This description obviously refers to mere diplomatic relations between these countries and Majapahit.
The imperial kingdom of Majapahit thus represents the final stage of a continuous process of state formation in pre-modern Indonesia. The state consisted of a series of concentric circles of authority. Its political control was strongest in its center, that is, the rajya of Majapahit. It decreased step wise in bhΰmijawa and the nusantara and ended up in mere diplomatic relations with the mandalita rastra or the "other countries" (desάntara) on Mainland Southeast Asia.
Despite the structural weakness of all pre-modern states of South and Southeast Asia, viz., the lack of actual political control outside the royal core region (rajya), under Majapahit two decisive structural changes had taken place. First, it was able to annex completely the neighboring kingdoms and little chieftaincies which usually still had surrounded the Stammland of the "Early Kingdoms" as autonomous watek or sάmantacakra. Majapahit thus created a considerably extended core region (rajya). Second, Majapahit succeeded in extending its uncontested hegemony over Java, to enforce tributary relations with a large number of outer islands and to establish diplomatic relations with kingdoms in Mainland Southeast Asia. Majapahit thus became Indonesia‘s first truly "Imperial Kingdom."
The concentric structure of the "empire" of Majapahit has indeed a strong resemblance to the conceptual model of the mandala state as described by O. W. Wolters and more recently by C. Higham in the context of Mainland Southeast Asia. [70] Derived from the ancient Indian Arthasastra, this concept is very suggestive and thus may be applied even more frequently to the pre-modern state in Southeast Asia. But while employing it in this context, one has to keep in mind that we have very little evidence that the term mandala was ever used in contemporary Southeast Asian sources in such a comprehensive way. As we have observed in the early Malay inscriptions of Srivijaya and in the Nagara-Kertagama, the political connotation of mandala always referred to a portion rather than to state as a whole. Particularly in the context of early kingdoms it denoted autonomous or semi-autonomous chieftaincies and principalities at the periphery of these states. As they were slowly integrated some of the mandala became provinces of the imperial kingdoms. In this later context, particularly in some of the great regional or imperial kingdoms of India (e.g. the Colas in South India and the Gaήgas in eastern India), [71] the term mandala was also used for provinces in the extended central core region. In order to avoid terminological misunderstandings, we have, therefore, to distinguish clearly between the ancient Indian concept of a mandala state system and the—rather different—medieval epigraphical meaning of the term mandala, denoting peripheral principalities or provinces. Moreover, as shown elsewhere, "the mandala concept does not give sufficient scope to structural changes which constitute the difference between the Early and the Imperial Kingdoms” [72]
[1] H. Kulke, ‘The Early and the Imperial Kingdom in Southeast Asian History," in: Southeast Asia in the 9th to 14th Centuries, ed. D.G. Marr and A.C. Milne (Singapore: Institute of Southeast Asian Studies, 1986), pp. 1-22; also "Indian Colonies, Indianization or Cultural Convergence? Reflections on the Changing Image of India‘s Role in South-East Asia," in: Onderzoek in Zuidoost-Azie. Agenda‘s voordejaren Negentig, ed. H. Schulte Nordholt (= Semaian, 3), Rijksuniversiteit te Leiden 1990, pp. 8-32.
[2] For further theoretical discussions, see D. Lombard, "Le concept d‘empire en Asie du Sud-Est," in: he Concept d‘Empire, ed. M. Duverger (Paris, 1980), pp. 433-41; see O.W. Wolters, History, Culture, and Region in Southeast Asian Perspectives (Singapore: Institute of Southeast Asian Studies, 1982); L. Gessick, ed., Centers, Symbols and Hierarchies: Essays on the Classical States of Southeast Asia (New Haven: Yale University Southeast Asian Studies
[3] B. Ch. Chhabra, Expansion of Indo-Aryan Culture During Pallava Rule (New Delhi, 1965) pp. 85
[4] See also F. H. van Naerssen, The Economic and Administrative History of Early Indonesia (Leiden/Koln, 977), pp. 18-23.
[5] See also J. G. de Casparis, "Some Notes on the Oldest Inscriptions of Indonesia” in: A Man of Indonesian Letters. Essays in Honour of Professor A. Teeuw, ed. C.M.S. Hellwig and S. O. Robson (Leiden, 1986), pp. 242-55 (= VKI, 121).
[6] H. Kulke, The Devaraja Cult (Ithaca: Cornell University Southeast Asia Program Data Paper 108,1979).
[7] H. B. Sarkar, Corpus of Inscriptions of Java, vol. I (Calcutta: Mukhopadhyay, 1971), pp. 15-2
[8] R. C. Majumdar, The Inscriptions of Champa (Lahore, 1927), pp. 1-
[9] G. Coedes, The Indianized States of Southeast Asia (Honolulu: University of Hawaii Press, 1968), p. 40.
[10] D. C. Sircar, Select Inscriptions, vol. I (Delhi: Motilal Banarsidass, 1965), p. 54, note
[11] Chhabra, Expansion, pp. 93-97.
[12] In his Ci-arutan Rock Inscription, a pair of human footprints "which belongs to the illustrious Purnavarman, the Lord of Tarumanagara" are compared with Visnu‘s footprints, whereas in the Kebon-Kopi Rock Inscription the footprints of his royal elephant are compared with those of Indra‘s elephant Airavata. (Chhabra, Expansion, pp. 93 and 95.)
[13] Wheatley, Nagara and Commandery, p. 233.
[14] P.-Y. Manguin, "Palembang et Sriwijaya: anciennes hypotheses, recherches nouvelles," BEFEO 76 (1987): 337- 402.
[15] From Palembang, the Kedukan Bukit and Talang Tuwo inscriptions were published by G. Coedes, "Les inscriptions malaises de ςrivijaya," BEFEO 30 (1930): 29-80; for the Telaga Batu [Sabokingking] inscription see J.G. de Casparis, Prasasti Indonesia, vol. II, (Bandung: Masa Baru, 1956), pp. 15-46; The more or less identical "mandala inscriptions" are known from Karang Brahi (Jambi) and Kota Kapur (Bangka), see Coedes, "Les inscriptions," and from Palas Pasemah (South Lampung), see Boechari, "An Old Malay Inscription of Srivijaya at Palas Pasemah," Pra Seminar Penelitian Sriwijaya (Jakarta, 1979), pp. 19-40; another badly weathered version of the mandala inscription was also found in the Lampung district and has been dealt with by Boechari in his article "New Investigations on the Kedukan Bukit Inscription" Untuk Bapak Guru (Bernet Kempers Festschήft) (Jakarta, 1986), pp. 33-56.
[16] de Casparis, Prasasti Indonesia, vol II, pp. 1-15.
[17] Coedes, "Les inscriptions," note 7 pp. 81-85; O. W. Wolters, Early Indonesian Commerce. A Study of the Origins of Srivijaya (Ithaca: Cornell University Press, 1967), pp. 15-29.
[18] de Casparis, Prasasti Indonesia, vol.11, p. 36, note 1.
[19] The translation follows, in most cases, ibid., pp. 36ff.
[20] J. G. de Casparis (ibid., p.37, n.7) calls his translation of haji pratyaya as "confidents of the king" as conjectural ("but we do not see another alternative"). Perhaps one could also think of the group of officers who are so frequently mentioned in later Javanese inscriptions as mangilala drazvya haji ("persons who collect the lord‘s property"); pratyaya has the meaning of "revenue, income or tax"; D. C. Sircar, Indian Epigraphical Glossary (Delhi, 1966), p. 262.
[21] de Casparis, Prasasti Indonesia, vol II, p. 21.
[22] In the Talang Tuwo inscription of 684 C.E. king Jayanasa is also addressed with the priestly (?) title (da)punta hiyam.
[23] H. Kulke, "Kadatuan Srivijaya—Empire or Kraton of Srivijaya? A Reassessment of the Epigraphical Evidence” in: The Ancient Southeast Asian City and State, ed. J. Stargardt (in press).
[24] J. G. de Casparis and Coedes translate kadatuan as "Empire" and "province" respectively. Boechari prefers "kingdom," although he admits that, strictly speaking, kadatuan is the equivalent of keraton, "Old Malay Inscription," p. 23.
[25] J. G. de Casparis (Prasasti Indonesia, Vol II, p. 14) and Boechari ( O l d Malay Inscription, p. 22) translate vanua as "country" and Coedes as "le pays" ("Les Inscriptions," p. 35, n. 12).
[26] de Casparis, Prasasti Indonesia, Vol. II, p. 14.
[27] O. W. Wolters, "Restudying Some Chinese Writings on Sriwijaya," Indonesia 42 (1986): 1-41.
[28] datu Srivijaya is mentioned in all mandala inscriptions, whereas vanua Srivijaya is known from the Kedukan Bukit inscription.
[29] Thus samaryyada is not mentioned in D. C. Sircar‘s Glossary.
[30] de Casparis, Prasasti Indonesia, Vol II, p. 43.
[31] Kulke, ‘The Early and the Imperial Kingdom."
[32] Thus, in the early Middle Ages, Tosali, the center of coastal Orissa in Eastern India, was surrounded by more than half a dozen such semi-autonomous mandala-principalities; see S.N. Rajaguru, Inscriptions of Orissa, vol. I, 2. [300-700 A.D.] (Bhubaneswar, 1958); and B. Misra, Dynasties of Medieval Orissa (Calcutta, 1933).
[33] Boechari, O l d Malay Inscription," p. 38 and Coedes, "Les inscriptions," p. 47.
[34] de Casparis, Prasasti Indonesia, Vol. II, pp. 4ff (particularly inscriptions a and b).
[35] In this connection further excavations at the newly discovered sites at Karanganyar in western Palembang will be most important. See P.-Y. Manguin, "Palembang et Sriwijaya."
[36] Sarkar, Corpus, pp. 15-24.
[37] Ibid., pp. 25-33.
[38] Ibid., vol. II, p. 68 (Mantyasih I copper-plates, B, 8).
[39] L. L. Gopal, "Samanta—Its Varying Significance in Ancient India," Journal of the Royal Asiatic Society of Great Britain and Ireland (1963): 21-3
[40] See above note 38.
[41] F.D.K. Bosch, "Het Lingga-Heiligdom van Dinaja," Tijdschήβ voor Indische Tool-, Land-, en Volkenkunde 64 (1924): 225-91; J.G. de Casparis, "Nogmals de Sanskrit-inscriptie op den steen van Dinojo," Ibid. 81 (1941) 499- 513; W.J. van der Meulen, ‘The Purϊ Putikesvarapavita and the Pura Kanjuruhan," in Bijdragen tot de Tool-, and- en Volkenkunde van Nederlandsch-Indie 132 (1976): 445-
[42] Coedes, "Les inscriptions” pp. 87-93,107-9; J. G. de Casparis, Prasasti Indonesia II, pp. 288
[43] See for instance de Casparis‘ detailed epigraphical study of the Tjaηdi Perot inscriptions of 850 C.E. in ibid., pp. 211-43 or W. F. Stutterheim‘s study of the Cunggrang II inscription of the year 929 C.E. in Tijdschrift voor Indische Taal-, Land-, en Volkenkunde 65 (1925): 208-θl.
[44] van Naerssen, Economic and Administrative History, pp. 46ff; Boechari, "Some Considerations of the Problem of the Shift of Mataram‘s Centre of Government from Central to East Java in the 10th Century A.D.," Bulletin of Research Centre of Archaeology of Indonesia 10 (Jakarta, 1976); J. Wisseman-Christie, "Raja and Rama: The Classical
[45] Inscription of Belingavan (Singasari), Sarkar, Corpus, vol. I, pp. 295-303.
[46] Inscription of Sugih Manek (Singasari), Sarkar, Corpus, vol. II, pp. 145-60.
[47] van Naerssen, Economic and Administrative History, pp. 53ff.
[48] Inscription of Gandasuli, line 8D; de Casparis, Prasasti Indonesia, vol. I, p. 61.
[49] J. G. de Casparis suggests "landstreek" (ibid., p. 68).
[50] Inscription of Kelurak, Sarkar, Corpus, vol. II, pp. 41-48.
[51] Copper-plates of Vuatan Tija (Manggung), Sarkar, Corpus, vol. I, pp. 250-61.
[52] Sarkar, Corpus, vol. II, pp. 144-60.
[53] Ibid., pp. 227-48.
[54] OJO, XLVIII and LI; see also W. J. van der Meulen, "King Sanjaya and his Successors," in: Indonesia 28 (1979): 17-54 (esp. pp. 24ff).
[55] Sarkar, Corpus, vol. II, p . 165.
[56] Boechari, "Rakryan Mahamantri i Hino. A Study on the Highest Court Dignitary of Ancient Java up to the 13th Century A.D.," Beberapa Karya Dcdam Ilmu-Ilmu Sastra (Publikasi Ilmiah, No. 2), Universitas Indonesia 1975/76, pp. 61-114.
[57] OJO,LXI.
[58] Kulke, ‘The Early and the Imperial Kingdom."
[59] See Boechari, "Sri Maharaja Mapanji Garasakan," Madjalah Ilmu-Umu Sastra Indonesia 4,1/2 (1968): 1-26.
[60] Coedes, Indianized States, p. 198.
[61] Nagara-Kertagama, 44,1; in Th. G. Pigeaud, Java in the 14th Century. A Study in Cultural History. The Nagara-Kertagama by Rakawi Prapaήca of Majapahit, 1365 A.Ό., 5 vols. (The Hague: Nijhoff, 1960).
[62] Nagara-Kertagama, 6,4.
[63] Kulke, ‘The Early and the Imperial Kingdom," p. 8ff.
[64] For the concept of the segmentary state see B. Stein, ‘The Segmentary State in South Indian History," in ealm and Region in Traditional India, ed. R. G. Fox (New Delhi, 1977), pp. 3-51; see also Subrahmanyam, "Aspects f State Formation."
[65] Nawanatya, 9a, in Pigeaud, Java in the 14th Century., vol. Ill, p. 121.
[66] Boechari, "Epigraphic Evidence on Kingship in Ancient Java” Madjcdah Emu-Ilmu Sastra Indonesia, 5,1 (1973): 119-26.
[67] Nagara-Kertagama, 15,3.
[68] Ibid., 15,1.
[69] Ibid., 12,6.
[70] See above note 2.
[71] R. Subbarayalu, ‘The Cola State," Studies in History (New Delhi) 4 (1982): 265-306; S.K. Panda, Herrschaβ und Verwaltung im δstlichen Indien unter den spaten Gangas (ca. 1030-1434) (Wiesbaden, 1986).
[72] Kulke, ‘The Early and the Imperial Kingdom” p. 13.
__________
Source: cip.cornell.edu