THE EARLY KINGDOMS 153

story with a clear and urgent purpose, the acceptance of which was an
important part of being Vietnamese. Have the Chams, Khmers, Burmans,
Tai-speakers, Malays, and Javanese made a similar use of history, or have
they been governed by entirely different notions of what history is 'for'?
With this question in mind, let us turn to narratives of Champa, Angkor,
Pagan, the early Tai kingdoms, Srlvijaya, and Majapahit.

CHAMPA
Champa is a generic term for the polities organized by Austronesian-
speaking peoples along what is now the central coast of Vietnam. Chinese
perceptions of these polities have been preserved and have ascribed
to Champa greater coherence and continuity than other evidence will
support. What is generally understood as Cham history is a twentieth-
century rationalization of scraps of evidence from inscriptions and Chinese
sources. The time has come to set this rationalization aside and to take a
fresh look. It immediately becomes apparent that the very concept of
Champa must be redefined. Rather than signifying a 'kingdom' in the
conventional sense of the word, Champa should more properly be under-
stood as an archipelagically-defined cultural-political space. Two clues
will provide entry into this unusual case, one geographical and one
cultural-linguistic.15
The land of Champa at its maximum extent stretched along the central
coast of what is now modern Vietnam from the Hoanh Son massif (Mui
Ron) in the north to Phan Thiet (Mui Ke Ga) in the south, a distance
of almost 1000 kilometres. Champa was comprised of small island-like
enclaves defined by the sea and the mountains. It was the closest that a
continental terrain could approximate the morphology of an archipelago.
The 'islands' were relatively isolated from the continent by a thick band of
mountains to the west, open to the sea on the east, and separated from
each other by lines of mountains that ran out into the sea. Bearing this in
mind, it is no accident that this is the one place (apart from the Malay
peninsula, which approximates a large island), where Malayo-Polynesian
peoples appropriated continental terrain.
The peoples of Champa, the lowland Chams and the upland Rhade and
Jarai, are ethnolinguistically Malay. Their organization of political space
can best be understood as a form of Malayo-Polynesian polity, quite
different from the polities we are accustomed to find in continental settings
or even on densely-populated islands such as Java. Political authority in
traditional Malayo-Polynesian culture grew out of maritime nomadism; it
was accordingly dispersed, with a preference for small groups enjoying
relative freedom to move about as they pleased. The land of Champa
offered opportunities for this type of organization, being broken up into
many small coastal enclaves with an extended mountainous hinterland.
Champa can be best understood as a kind of archipelago where ambi-
tious leaders repeatedly established centres of authority, simultaneously
15
My understanding of Champa and of new directions in Cham studies is indebted to
discussions with Nora Taylor. She bears no responsibility for my errors.

Cambridge Histories Online © Cambridge University Press, 2008

154 FROM PREHISTORY TO c. 1500

and in different places. In lieu of constructing a schematic and meaningless

'narrative' of Cham 'history', I propose to look at five regions that may be
understood as 'island-clusters' within the larger archipelago. The evidence
sits more comfortably in such a framework than it does in the conventional
category of a unified kingdom.
Beginning in the north, Quang Binh, Quang Tri and Thua Thien
Provinces correspond with what the Chinese called Lin-yi. Lin-yi, as some-
thing in the vicinity of Hue, held the attention of Chinese record-keepers
from the late second century to the mid-fifth century, after which the
Chinese appear to have applied the name to something further south in
the vicinity of Da Nang. The outstanding feature of Lin-yi from the Chinese
point of view was that it appeared out of the debris of the crumbling
southern border of the Han empire and was a source of incessant frontier
raids until 446 when a Sino-Vietnamese expedition destroyed its centre at
Hue. The Chinese continued to apply the name Lin-yi to something they
perceived further south for another three centuries. Instead of presuming
that 'the Chams moved their capital further south', it appears more likely
that whatever had been happening in this region was in some unknown
way altered after 446 and that regions further south simply continued
along lines of development already established.
The so-called Lin-yi that existed from the late second century to 446
cannot be verified with architectural or inscriptional evidence. According
to the Chinese sources, this region was the southernmost frontier jurisdic-
tion of the Han empire for three centuries before someone identified as a
son of a local Han magistrate proclaimed himself king in the waning years
of the Han dynasty. For the next two and a half centuries, the dominant
feature of this Lin-yi is that its kings appear to have defined their ambitions
in response to the rhythms of dynastic power in China; they seem to have
been preoccupied with the Sino-Vietnamese frontier and with opportu-
nities for plunder and expansion beyond that frontier. This perception can
be attributed to the priorities of Chinese annalists. In fact, warfare along
the border during this time increased in times of Chinese dynastic
strength, when Sino-Vietnamese armies were most active, and decreased
when Chinese dynasties were weak and least able to project military
power into the far south. 16 This suggests that warfare may have been as
much a factor of initiative from the north as from the south. The border
was unstable and powers both in the north and in the south endeavoured
to maximize their control of terrain. The modern territories of Quang Binh,
Quang Tri and Thua Thien would in fact be a contested frontier zone until
the fifteenth century when the arena of contention shifted further south.
After 446, Sino-Vietnamese annalists perceived kings of polities beyond
the southern border as being situated in the vicinity of the modern city of
Da Nang, in the modern province of Quang Nam. These annalists stopped
using the name Lin-yi after 758, and from 875 began to use the name Chan-
ch'eng, understood as equivalent to Champapura or 'City of Champa'.
This region is rich in architecture, statuary, and inscriptions in Sanskrit
and Cham. The major archaeological sites are Mi-son, Dong Duong, and

16
K. W. Taylor, The Birth of Vietnam, Berkeley, 1983, 106-9, 115-18.

Cambridge Histories Online © Cambridge University Press, 2008

 THE EARLY KINGDOMS 155

Tra Kieu, each of which represents a distinctive artistic tradition. 17 The

name Amaravati has been applied to some of the artistic remains by French
scholars on the basis of a presumed relationship with the Indian style of
that name. The name Indrapura has also been applied to this region. The
task of absorbing the vast amount of archaeological evidence already
found and still being found has barely begun. Inscriptions reveal a rhythm
of political life centred on an aspiring leader's ability to erect a linga and to
protect it from rivals.18 According to the usual reading of the Sino-
Vietnamese annalists, Cham-Viet warfare in the late tenth century led to
the abandonment of the Quang Nam region as an arena for aspiring Cham
kings, though in fact the annalists can be read as being ambiguous on that
point.19 This coincided with the appearance of a new Vietnamese kingdom
separate from the Chinese imperial system and may suggest that the
process by which the Vietnamese established their position of separation
with regard to China also involved the assertion of a greater measure of
military ascendancy in the south. However, inscriptional evidence reveals
that Cham kings continued to be active in this region into the late twelfth
century,20 and the Sino-Vietnamese annalists themselves do not deny that
this territory belonged to Cham kings until the late fifteenth century.
From the eleventh to the fifteenth centuries, Vietnamese annalists locate
the kings they call Cham in the vicinity of the modern city of Qui Nhon.
The Cham capital, often identified by the name Vijaya, was twice sacked
by Vietnamese seaborne armies in the eleventh century as part of the
process by which the Vietnamese were developing their dynastic space
and exploring the limits of their maritime power. In the twelfth and early
thirteenth centuries, Khmer armies, with Angkor at the peak of its power,
repeatedly occupied parts of the Cham territories and endeavoured to
establish an overlordship over the Cham kings; Cham kings remained
active, however, and even managed to sack Angkor in 1177. In the late
thirteenth century, Chams and Viets allied against the Mongol-Yuan
invasions; the momentum of this alliance eventually led to the visit of a
Vietnamese king to Champa in the early fourteenth century, and the
marriage of a Vietnamese princess to the Cham king. According to Viet-
namese annalists, however, territorial quarrels over the region of Quang
Binh, Quang Tri and Thua Thien eventually led to hostilities, and, during
the era of Tran dynastic decline in the late fourteenth century, as we have
seen, an able Cham king, known to the Vietnamese as Che Bong Nga,
repeatedly invaded and plundered the Vietnamese lands, thrice sacking
the Vietnamese capital. Following the death of Che Bong Nga in 1390, Ho
Quy Ly, and then the fifteenth-century Le kings, presided over new
assertions of Vietnamese power in the Cham territories. In 1471, according
to Vietnamese annalists, a Vietnamese army seized the Cham capital at
Vijaya and the Vietnamese annexed everything north of what is today the
southern border of Binh Dinh Province.
17
Jean Boisselier, La Statuaire Du Champa, Paris, 1963.
18
L. Finot, 'Notes d'epigraphie: les inscriptions de Mi-son', BEFEO, 4 (1904), inscriptions
no. 4, 12, 14, & 15.
" G. Ccedes, The Iiidianized States of Southeast Asia, ed. Walter Vella, trans. Susan Brown
Cowing, Honolulu, 1968, 124-5; Ngo Si Lien, ban ky 1, 189-94 passim.
20
Finot, 'Les inscriptions de Mi-son', inscription no. 25.

Cambridge Histories Online © Cambridge University Press, 2008

156 FROM PREHISTORY TO c. 1500

The status of the region of Q u a n g Binh, Q u a n g Tri and Thua Thien

during the t h o u s a nd years after the so-called kings of Lin-yi ceased to rule
there in the fifth century and until an unambiguous and final Vietnamese
authority appears to have been established in the fifteenth century has yet
to be carefully studied. It was clearly a border zone in which Viet and
C h a m peoples mingled. The Vietnamese annalists claim that northern
portions of this region were annexed in the eleventh century and southern
portions were obtained by marriage alliance in the early fourteenth centu-
ry. But Ho Q u y Ly's campaigns against Champa at the turn of the fifteenth
century reveal that this was still a contested area. Similarly, the status of
the Q u a n g N a m region after the tenth century and until the Vietnamese
annexation of 1471, has yet to be clearly understood. This was recognized
by the Vietnamese as Cham territory, and Cham kings continued to rule
from there, some of them emplaced by Khmer armies, but overall it
appears to have been a less propitious place for Cham kingship after the
tenth century than it had been before.
This should not lead to the conclusion that the Q u a n g Binh, Q u a n g Tri
a n d Thua Thien region after the fifth century and the Q u a n g Na m region
after the tenth century were simply absorbed by Cham polities further
south. O n the contrary, it is more likely that, exceptional periods of Cham
leadership aside, such as the case of Che Bong Nga, regional leaders
exercised a kind of a u t o n o m o us authority appropriate to local circum-
stances. The information from the Vietnamese annals about Cham-Viet
relations must be treated carefully, for surely it reflects presumptions
about interstate relations derived from experience with Vietnam's north-
ern neighbour. Vietnamese-style historical narrative was not adequate to
describe the relatively diffuse, personalized kind of authority that must
have characterized C h a m political experience.
The Vietnamese annexation of the coast d o w n to Cu-mong in the late
fifteenth century was not the end of Champa. A fourth Cham region
centred in the vicinity of the modern city of Nha Trang, in the modern
province of Khanh Hoa, had been the home of kings from the beginning of
Cham history and has archaeological and inscriptional evidence com-
parable to the Q u a n g N am region in terms of chronology, quantity,
and sophistication. Scholars sometimes apply the name Kauthara to this
region. Sino-Vietnamese annalists recorded what appears to have been a
perception of this region in the eighth and ninth centuries u n d e r the name
H u a n - w a n g . C h a m kings ruled here until the end of the seventeenth
century. Thereafter, Cham kings continued to rule under Vietnamese
overlordship in what is the moder n province of Thuan Hai until 1832. 21 In
this region, often referred to by scholars as Panduranga, Cham kings had
been ruling for centuries. The majority of C h a m s living in Vietnam today
are in Thuan Hai. 2 2
O n e of the w o n d e r s of Cham history is that a Malayo-Polynesian people
was able to compete for space in a continental environment for so long.
The archipelagic, maritime nature of this continental terrain is one way to

21
Po Dharma, Le-Panduranga (Campa) 1802-1835, ses rapports avec le Vietnam, Paris, 1987.
22
Vien khoa hoc xa hoi thanh pho Ho Chi Minh, Nguoi Cham o Thuan Hai, Thuan Hai, 1989.

Cambridge Histories Online © Cambridge University Press, 2008

 THE EARLY KINGDOMS 157

understand this. Another dimension that deserves further consideration is

the participation of the upland peoples in the Cham story. Temples and
other archaeological remains can be found in the mountains, at least one
king ruled from the mountains, and there is evidence of close relations
between upland and coastal leadership groups. Much scholarly attention
has been spent on lists and genealogies of kings as metaphors for a
kingdom, but now it is clear that there were many kings ruling simulta-
neously in different places. Who were these kings? They were never called
Cham; rather, they were called kings of Champa. It is in fact difficult to
get a clear sense of who the Chams were. Judging from surviving Cham
populations, we are encouraged to speak of there being several kinds of
Cham peoples. And it is increasingly clear that the participation of peoples
from the mountains, the Rhade and the Jarai, as soldiers and even kings, is
more than a possibility. Furthermore, the significance of Champa as a
network, or series of networks, of ethnic, religious, political, and commer-
cial relationships connecting the Cham territories with the Malay world of
peninsular and insular Southeast Asia is still poorly understood. There
is in fact much evidence from Champa that has yet to be studied; what is
meant by Champa appears to be on the threshold of a major revision.23
The construction of Cham history will have a significant impact upon
the development of Vietnamese historiography, which is at present in a
preliminary phase of making space for Cham voices to be heard in the
earlier history of the terrain now part of Vietnam.

ANGKOR
Turning to the lower Mekong basin, we find a very different historical
experience from that of the Viets or the Chams. Here there was no
experience with the soldiers and officials of a neighbouring empire, nor the
awareness of boundaries, in terrain and in culture, that such an experience
produced among the Viets and Chams. Information about the outside
world available to Khmer leaders arrived as news about Hindu gods and
forms of Hindu and Buddhist devotion as well as cosmological notions of
political space that were expounded in the Sanskrit language. This news
arrived, we assume, chiefly by way of maritime trade, and encouraged
openness and receptivity toward distant contacts and awareness of being
part of a borderless world. Early Khmer leaders learned to justify their
authority by placing it in a universal context of devotion that could
fully absorb the religious aspirations and compel the loyalty of their
followers. In a process of developing the theory and practice of an
increasingly centralized political space, warfare among rival hegemons was
rationalized as corresponding to deified moral conflict on a universal scale.
The emergence of the Angkor polity from the ninth century represents the
accumulated political and cultural wisdom from generations of efforts to
organize a political order in a relatively diffuse socio-economic environment.
Unlike the Viets, who developed their polity in a relatively confined
23
Finot, 'Les inscriptions de Mi-son', inscription no. 21. T. Quach-Langlet, 'Le cadre geogra-
phique de l'ancien Campa', in Actes du seminaire sur le Campa, Paris, 1988, 36.

Cambridge Histories Online © Cambridge University Press, 2008

158 FROM PREHISTORY TO c. 1500

locale that was densely populated from an early time, and the Chams,
w h o s e ambitions were for the most part defined by coastal enclaves, other
peoples in Southeast Asia inhabited more expansive landscapes where
h u m a n settlement remained an important variable for a longer time. The
Khmers inhabited the lower Mekong basin, which included the Tonle Sap,
or Great Lake, a natural reservoir for the annual monsoon floods of the
Mekong. During the earliest, or pre-Angkorean, centuries of Khmer histo-
ry, there was n o fixed centre, nor can it even be said that there was a single
Khmer polity. Khmer leaders strove to promote and enforce their authority
by demonstrations of battlefield and devotional prowess, the effectiveness
of which seldom h a d any e n d u r i ng value beyond their individual life-
times. 2 4 However, in the ninth century, Khmer political life became
centred at Angkor. Angkor is located near the northwest shore of the
Tonle Sap, with good water transport from all the ricefields in the drainage
basin of that lake as well as the lower Mekong plain. At the same time,
it is well situated for land contact with the basins of the M u n and Chao
Phraya, w h e r e more ricefields could be found. Once Khmer settlement
had exploited the rice-growing potential of this region to a minimally-
necessary level, and once Khmer leaders found ways of organizing their
authority over much of this region, Angkor was the favoured site as long
as agriculture remained the primary source of wealth. Pre-Angkorean and
post-Angkorean Khmer centres were located to the east and south, along
the Mekong with direct access to the sea and the commerce-generated
wealth that this access afforded; but Angkor d e p e n d e d upon rice.
The establishment of the Angkorean polity is associated with the career
of Jayavarman II during the first half of the ninth century. Prior to this
time, there was a multiplicity of polities in the lower Mekong basin,
relatively small a n d transitory realms representing the personal achieve-
ments of particular individuals rather than institutionalized political
systems. Chinese record-keepers organized their information about these
polities in the shape of two successive kingdoms, which they called Fu-
nan, from the second to the sixth centuries, and Chen-la, from the sixth
through to the eighth centuries. The Chinese perception of a change from
'Fu-nan' to 'Chen-la' in the sixth century appears to correspond to a
transition from coastal or riverine entrepots linked to the trade route
between India a n d China to a more inland focus u p o n ricefields.
The earliest k n o w n maritime trade route between India and China
followed the coasts, except for land transit across the Kra isthmus and
transit through the natural and man-mad e channels of the lower Mekong
plain. The archaeological site of Oc-eo, in southern Vietnam near the
Cambodian border, appears to have been situated at a strategic junction
of canals that linked the Gulf of Siam with the main channels of the
Mekong. 2 5 Oc-eo w a s an entrepot from the second to the sixth centuries,
and is generally associated with the 'Fu-nan' era of pre-Angkorean history.
But, beginning in the fourth century, an all-sea route was pioneered
24
O. W. Wolters, 'Khmer "Hinduism" in the seventh century', in R. B. Smith and W. Watson,
eds, Early South East Asia: Essays in Archaeology, History and Historical Geography, London and
New York, 1979.
25
Paul Wheatley, Nagara and Commandery, Chicago, 1983, 119-46.

Cambridge Histories Online © Cambridge University Press, 2008