The Khmer Empire

A region that had to exist

formationDrafted from Higham, The Civilization of Angkor (2001)

Angkor is a rice state on an inland plain, and that fact separates it sharply from the maritime polities of the archipelago — Srivijaya, Melaka, Aceh, Ternate — whose power ran on controlling straits rather than territory. Angkor built at a scale nothing in the islands approached, and for most of its six centuries the sea route that made those port states rich was something that happened to other people. Pagan on the Irrawaddy, Dai Viet on the Red River and Ayutthaya on the Chao Phraya are the same kind of object: inland agrarian states whose wealth came from taxing a floodplain, not a shipping lane. The islands and the mainland peninsula were not one Southeast Asian world with a size difference between them; they were two different economies.

What the archaeology changed

Coedès called these mainland and island polities alike Indianized states, and the framework is built into that name: temple, script, Sanskrit, kingship and law arriving from the subcontinent and being received by societies that had comparatively little of their own to set against them.

Higham's excavation of the pre-Angkorian Mekong is the main reason that reading no longer stands on its own. There were substantial, stratified, wealthy settlements on this ground for a thousand years before any Sanskrit inscription, with their own bronze industry, their own trade and their own hierarchies. What arrived from India was a vocabulary in which existing power could describe itself, adopted by people who already had the power.

Higham, The Civilization of Angkor (2001) · PsamatheM · CC BY-SA 4.0

The water

expansionDrafted from Higham, The Civilization of Angkor (2001)

Angkor's plain has a monsoon that delivers most of the year's rain in five months and almost none in the rest, and a lake — the Tonle Sap — that reverses the flow of its own outlet river twice a year and swells to five times its dry season size.

The city built for that. The West Baray, finished in the eleventh century, is a rectangular reservoir eight kilometres long and two wide, held by earth embankments, and it is one of four on the plain. Around them ran a system of channels, moats and dykes that the aerial and lidar surveys of the last few decades have shown extending across roughly a thousand square kilometres — a low-density urban landscape of house mounds, ponds and rice fields with the temples inside it, and a population that no other pre-industrial settlement in Southeast Asia came close to.

Higham, The Civilization of Angkor (2001) · Vladimir.ira · CC BY-SA 3.0

What the barays were for has been argued about for a century: irrigation storage, or ritual replicas of the cosmic ocean, or both, or flood management. The argument matters less than the fact that whatever they were for, an enormous quantity of directed labour built them — and the inscriptions say where that labour came from.

Rice, not tolls

Angkor lived on land and on labour, and on essentially nothing else. Against the rest of Southeast Asia that is worth pausing on.

The Khmer temple foundation inscriptions are, functionally, tax records written from the receiving end. They list what a foundation was given: named villages, areas of rice land, and people — counted, by category, by the hundred. Read them and the state's income resolves into rice and work, taken from cultivators through the officials who held the assignment.

Now compare the polities of the straits a thousand miles south. Srivijaya and Melaka lived on tolls almost exclusively; they taxed movement rather than production, and their power ended where the water did. Angkor was the largest state anywhere near them and took nothing from trade worth recording.

The two kinds of state are not variations on a theme. One is a machine for organising a floodplain and the people on it; the other is a gate. They produce different capitals, different armies, different reasons to fight and different ways of dying.

The city founded twice

expansionDrafted from Cowing, The Indianized States of Southeast Asia (1968)

Jayavarman II's own two immediate successors left almost no epigraphic trace, and it was Indravarman, who took power in 877 with no clear blood tie to either of them, who resumed the inscriptional record his spiritual teacher's own family had kept since the founding. In his first year he began the Indratataka, a reservoir north of the capital at Hariharalaya - a reservoir built, on Coedès's reading, for both irrigation and ritual purposes, and the first of a pattern of larger and larger water-works every strong king after him would repeat. In 881 he raised Bakong, the kingdom's first great temple built in stone, for a royal linga named - by the custom of pairing a god's name with the founder's own - Indresvara.

Cowing, The Indianized States of Southeast Asia (1968) · Jean-Pierre Dalbéra · CC BY 2.0

A king who moved the whole city

Indravarman's son Yasovarman, who succeeded him in 889, inherited both a peaceful kingdom and a father's habit of building on an ever-larger scale, and within his first years in power he did what no Khmer king had done before him: he moved the capital outright. The inscription of Sdok Kak Thom records it plainly - "the king founded the city of Yasodharapura and took the Devaraja away from Hariharalaya to establish it in this capital." He raised the temple mountain of Phnom Bakheng at the new city's center, for a linga called Yasodharesvara after himself, and dug a reservoir seven kilometers long north of the new capital, the Yasodharatataka - what modern maps call the Eastern Baray - bordered by steles recording his own genealogy and the scale of his own ambition in Sanskrit verse. Around the reservoir he settled monasteries for every sect his own religious eclecticism could accommodate, an urban foundation built, in effect, all at once around a king determined to outbuild his own father.

Two obscure reigns, and a city abandoned

Yasovarman's own sons who followed him, Harshavarman I and Ishanavarman II, left little behind them - a small temple at the foot of Phnom Bakheng, a handful of inscriptions, posthumous names and almost nothing else. What happened next was, by Coedès's own reading, a usurpation dressed in family language: an uncle of the two brothers left Yasodharapura around 921, "taking the Devaraja with him," and built an entirely new capital at Koh Ker under the name Jayavarman IV. There he raised a five-stepped pyramid crowned by a linga the inscriptions call, in an escalation of the usual formula, "the god who is the royalty" - a monument on a scale that dwarfed anything built at Yasodharapura itself. The new capital did not outlast its builder by much. About twenty years after it was founded, Jayavarman IV's own line gave way to a cousin's, and the court abandoned Koh Ker for the city Yasovarman had built, which had never stopped being, in every sense but the current address of the king, Angkor's real home.

Cowing, The Indianized States of Southeast Asia (1968) · Lohen11 · CC BY 4.0

A city restored

expansionDrafted from Cowing, The Indianized States of Southeast Asia (1968)

Rajendravarman came to power around 944 through a claim built on two lines at once - nephew to Jayavarman IV through his father, and grandson of Yasovarman I's own sister through his mother - a pedigree strong enough to override the more direct but still very young heirs of Jayavarman IV's own house. His first act as king was to reverse his predecessors' relocation outright: he brought the Devaraja back to Yasodharapura and rebuilt the palace there, an inscription comparing the work to Kusha's own restoration of Ayodhya in the Ramayana - a return the court's own poets framed as putting things back the way they belonged.

A raid on Champa's own goddess

Rajendravarman's reign is otherwise known to history mostly through what his own inscriptions boast of rather than narrate in detail: his brilliance, one text says, "burned the enemy kingdoms beginning with Champa." Around 950 he sent an army into Cham territory that reached the temple of Po Nagar at Nha Trang and carried off its gold cult statue - a raid remembered on the Cham side of the same conflict as a genuine desecration, and on the Khmer side as proof of a king who could reach as far as his predecessors ever had.

A temple pink stone made famous

Rajendravarman died in 968, and near the very end of his reign a grandson of Harshavarman I, serving as a court teacher, founded a temple that would outlast the memory of nearly everyone connected to it: Banteay Srei, built of a fine-grained pink sandstone that let its carvers work detail no earlier Khmer monument had attempted. Rajendravarman's own son, Jayavarman V, inherited the throne as a small child and did not finish his own education until 974; his long but thinly documented reign built a new royal residence and held the kingdom together without the wars or usurpations that had marked three of the previous four successions - a period, by the standard of what came immediately before and after it, of simple continuity.

Cowing, The Indianized States of Southeast Asia (1968) · Olaf Tausch · CC BY 3.0

Three kings, one throne

expansionDrafted from Cowing, The Indianized States of Southeast Asia (1968)

For most of the first decade of the eleventh century, Khmer inscriptions record three kings at once. Udayadityavarman I, a nephew of Jayavarman V, took the throne in 1001; a rival, Jayaviravarman, was reigning at Yasodharapura itself from at least 1003; and a third prince, Suryavarman - unrelated to the Suryavarman who would take the same name over a century later - held territory to the east from 1002. Between 1005 and 1007 Suryavarman led a campaign the inscriptions describe in unusually candid terms: "he seized the kingdom from a king in the midst of a host of other kings," damaging sacred sites along the way. The war lasted nine years, and Suryavarman's own occupation of the capital dates to around 1010 - though his later inscriptions, asserting an unbroken legitimacy, backdated his own accession to 1002, the year his first rival disappeared from the record.

A usurper who governed like a founder

Suryavarman I claimed descent from Indravarman on his mother's side and a marriage tie to Yasovarman's own line through his wife - claims Coedès treats as at least partly a legitimizing fiction laid over a victory won by force. Once secure, he ruled like a builder rather than a conqueror still fighting for his throne: he completed the Phimeanakas and Ta Keo, raised entrance pavilions for the royal palace, and had officials swear a formal oath of allegiance to him carved into the palace walls in long lists of names. He favored Buddhism enough to earn the posthumous name Nirvanapada, without interrupting the state cult of the Devaraja his Shaivite predecessors had maintained - the same syncretism that let a usurper who leaned Buddhist still keep the priestly family that had served the Devaraja since Jayavarman II's own founding.

A gift sent south, against a threat from the sea

Around 1012, feeling threatened by a rival power, Suryavarman I sent a chariot as a gift to the south Indian Chola king Rajendrachola I - a diplomatic overture Coedès reads as aimed at Srivijaya, whose own king at Kataha the Cholas would attack in a major naval expedition soon after. It was as close as an inland rice kingdom on the Cambodian plain ever came to entering the maritime politics of the straits a thousand miles to its south, and it came only because a usurper still consolidating his own throne needed friends wherever he could find them.

The mountain of gold, and the lake that outgrew the last one

Suryavarman I died in 1050 and was succeeded by Udayadityavarman II, who spent much of his own sixteen-year reign putting down revolts a general named Sangrama suppressed on his behalf. What he is remembered for instead is a single temple and a single reservoir: the Baphuon, a temple-mountain a Chinese visitor two centuries later would still call genuinely impressive, and the Western Baray - at eight kilometers by two, larger even than Yasovarman's own Eastern Baray, which by this point was already showing signs of silting up. It is this reservoir the water systems of Angkor are still measured against today.

Cowing, The Indianized States of Southeast Asia (1968) · Diego Delso · CC BY-SA 3.0

A funerary temple for a living king

expansionDrafted from Cowing, The Indianized States of Southeast Asia (1968)

Suryavarman II took the throne in 1113 the way several Khmer kings before him had: by killing a rival. He fought his uncle Dharanindravarman I in a single day of battle, and the inscriptions describe him leaping onto the head of his enemy's war elephant to strike him down in person. A second rival king fell with him. Suryavarman II had the court Brahmin who legitimised royal successions crown him the same year, folding two contested claims into one uncontested reign.

A soldier-king who mostly lost abroad

He spent much of the next three decades at war. Embassies to Song China opened within a few years of his accession, but the real business of the reign was fought closer to home, against Dai Viet to the northeast and Champa to the east. The results were mixed at best: a 20,000-man invasion of Vietnamese territory in 1128 was driven back, a fleet sent to raid the Vietnamese coast achieved little lasting effect, and a renewed campaign a decade later failed outright. Only against Champa did he manage a real conquest, seizing Vijaya in 1145 and occupying northern Cham territory for four years before a Cham king retook the capital and ended the occupation in 1149. A final expedition against Dai Viet the following year collapsed to disease and bad weather before it ever reached battle.

Cowing, The Indianized States of Southeast Asia (1968) · Michael Gunther · CC BY-SA 4.0

What he built instead

It was the building programme, not the wars, that outlasted him. Suryavarman II endowed temples across the kingdom's older Shaivite foundations, but the project that carries his name is Angkor Wat itself: constructed during his own lifetime, on a scale no Khmer temple before or since matched, and built from the start as his own funerary temple. Where earlier kings had dedicated their great foundations to Shiva, Suryavarman II built his to Vishnu, and after his death he was venerated there as a Vishnu incarnation under the posthumous name Paramavishnuloka — a shift in royal devotion that put Angkor's most famous monument on the same religious ground as the Vaishnavite courts of contemporary Java, rather than the Shaivite tradition Angkor's own kings had followed since Jayavarman II.

Cowing, The Indianized States of Southeast Asia (1968) · Charles J Sharp · CC BY 2.5

An ending nobody recorded

Nobody wrote down when or how Suryavarman II died. His last surviving inscription dates to 1145, but he was almost certainly still reigning when the disastrous 1150 campaign against Dai Viet was launched in his name. His successor, Dharanindravarman II, was not his son but a cousin — and a Buddhist, breaking a Hindu orthodoxy at court that had held, with only individual tolerance for Buddhism alongside it, since the kingdom's founding. Whether the succession passed peacefully or through the kind of palace violence that had put Suryavarman II himself on the throne, the silence in the inscriptions is the only trace left of it.

The Chams in the city

peakDrafted from Cowing, The Indianized States of Southeast Asia (1968)

In 1177 a Cham fleet came up the Mekong, crossed the Tonle Sap and sacked Angkor.

It is a startling reversal of the ordinary picture of the mainland, in which the great inland rice kingdoms are the powers and the coast is a margin. The largest land power in Southeast Asia, forty miles from the nearest navigable approach and a very long way from the sea, was taken from the water by a string of coastal principalities with a fraction of its population.

What followed made the polity's reputation. Jayavarman VII drove the Chams out, took the throne in 1181, invaded Cham territory and held parts of it for two decades, and then built. The Bayon, Ta Prohm, Preah Khan, Banteay Kdei, a hundred and two hospitals, a road network with rest houses at intervals — the largest construction programme in Khmer history, carried out in about thirty-five years, by a king who was already in his sixties when he began.

Musée Guimet, Paris · Cowing, The Indianized States of Southeast Asia (1968) · Vassil · Public domain

The Bayon's outer gallery reliefs show the naval battles against the Chams in detail: the boats, the rowers, the men going into the water, the fish below them. It is a monument to a war the polity had lost the first round of, carved by the side that won the second.

The conversion

Jayavarman VII made Mahayana Buddhism the state religion of a kingdom that had been Shaivite for three hundred and eighty years.

Cowing, The Indianized States of Southeast Asia (1968) · Dudva · CC BY-SA 3.0

Angkor never propagated a religion of its own past its own borders. It was on the receiving end every time: Shaivite from India, Mahayana in 1181, and Theravada from the Mon and Sinhalese world afterward, which is what the country has practised ever since. A polity can be the largest religious builder on a continent and still never be the source anyone else's faith travelled outward from.

Nothing on the scale of the Bayon programme was built at Angkor after Jayavarman VII died. His reign was simultaneously the kingdom's greatest exertion and the last of it.

The Thai script

aside?What aside meansDrafted from Cowing, The Indianized States of Southeast Asia (1968)

The Thai script devised at Sukhothai in the late thirteenth century takes Khmer letterforms and applies them to a language of a completely different family.

Khmer is Austroasiatic and is not tonal. Thai is tonal, and tone is not optional information in it — it distinguishes one word from another. So the adaptation had to invent signs for something the source script had no concept of, and it did, by pressing the Khmer consonant classes and a set of new diacritics into service.

That is a modification of a writing system rather than a borrowing of one, and it is the reason the Thai script is Angkor's inheritance rather than Sukhothai's invention from nothing.

And what it means here

Ayutthaya, the polity that destroyed Angkor, wrote in it.

When Ayutthaya took Angkor in 1431 it carried off the court ritual and the titulature with everything else, and it recorded those Angkorian titles in letters that were themselves Angkorian, reshaped a hundred and fifty years earlier by a different Thai kingdom.

What Angkor left behind, it left to the place that ended it.

The water that stopped

contractionDrafted from Cowing, The Indianized States of Southeast Asia (1968)

Angkor was a hydraulic city, and the thing about hydraulic cities is that they have to be maintained by the state that built them, for ever.

What Jayavarman VII spent

The reign that looks like the peak is also where the trouble starts.

Jayavarman VII, after retaking the capital from the Chams in 1181, built more than any Khmer king before or since: the Bayon, Ta Prohm, Preah Khan, Banteay Kdei, the new walled city of Angkor Thom, and — this is the part usually skipped — a network of a hundred and two hospitals and a hundred and twenty-one rest-houses along the roads, each endowed.

The Ta Prohm inscription counts the establishment: some eighty thousand people attached to that one foundation, including six hundred dancers, and villages supplying it. Preah Khan's is comparable.

Cowing, The Indianized States of Southeast Asia (1968) · Diego Delso · CC BY-SA 3.0

Every one of those is land and labour removed from the ordinary revenue and assigned to a religious foundation in perpetuity. It is the same instrument that was dismantling Pagan in the same century, at a larger scale, in a shorter time.

Then the religion changed

Jayavarman VII's foundations were Mahayana Buddhist. His successors were not, and the reaction was violent: the Buddha images in his temples were systematically defaced and recut as Hindu figures, which is visible on the stones today.

Then, over the thirteenth and fourteenth centuries, the country became Theravada — the monastic Buddhism of Sri Lanka, spreading through the villages rather than imposed from the court.

That mattered institutionally more than doctrinally. The whole Angkorian apparatus rested on the devaraja cult and on temple foundations built and staffed by a king to generate merit for a king. Theravada does not work that way: merit is personal, monasteries are modest, and a wooden monastery in a village does not require a state.

Zhou Daguan, the Chinese envoy who lived at Angkor in 1296 and left the only eyewitness account of it, describes the Theravada monks in yellow robes as simply the religious establishment of the place. Within a century of the last great temple, the ideology it had been built to express was gone.

The barays

The reservoirs are the other half.

The West Baray, the East Baray, the moats and the channels between them are an engineered landscape of enormous scale, and what they did is debated — irrigation in the dry season, flood control, water storage, ritual display, or all four. What is not debated is that a system like that silts up, and that keeping it clear requires organised labour every year, indefinitely.

The East Baray was dry by the time Zhou Daguan arrived. Sediment cores and the excavation of the channels show breaches, patches, and reworkings that get cruder over time — an infrastructure being repaired by people with fewer resources than the people who built it.

And then the climate. Tree-ring work from the region indicates severe decadal droughts in the fourteenth and early fifteenth centuries, broken by intense monsoons — which is the worst possible sequence for a canal network, because drought starves it and flood tears it apart.

The direction of the country

Underneath all of it, the centre of gravity was moving.

Maritime trade with China was growing through the fourteenth century, and it favoured places on rivers that reached the sea. Angkor is a hundred and fifty kilometres inland, on a plain optimised for rice and ritual rather than for shipping.

The Khmer court eventually moved south towards the Mekong and Phnom Penh, and the reasons are commercial as much as military. By the time the Siamese took Angkor in 1431 the capital had already been losing its argument for existing for a century, and what looks like a fall is mostly the last stage of a relocation.

1431

endDrafted from Cowing, The Indianized States of Southeast Asia (1968)

Ayutthaya took Angkor in 1431 and the Khmer court moved south, toward what became Phnom Penh, and did not come back.

And yet "conquest" is not quite the right word, which is worth saying plainly. Ponhea Yat, the last king to rule from Angkor, reigned for about thirty years after 1431. The dynasty was not extinguished, the country was not annexed, and Cambodia continued as a kingdom into the twentieth century. What ended in 1431 was the capital, and the six-century programme of building and irrigation that the capital was.

What was carried off

The Ayutthayan army took the brahmans.

That is not a figure of speech. Ritual specialists, the court's Sanskritic apparatus, the consecration rites, the palace vocabulary, the legal formulae — the working parts of Angkorian kingship went to the Chao Phraya, and Siam governed with them until the nineteenth century. The Bangkok dynasty was still using them after that.

It is a very rare thing for a state to acquire an entire court culture by capturing the city that had it. Ayutthaya afterwards asserted its own kingship in an idiom it had taken by force from the place it had just destroyed.

The temples, and the flag

afterlifeDrafted from Cowing, The Indianized States of Southeast Asia (1968)

Angkor was never lost, never rediscovered, and is on the national flag. Each of those needs saying because the first two are the opposite of what is usually believed.

The city that was not abandoned

The story that Henri Mouhot found Angkor Wat swallowed by jungle in 1860 is a piece of nineteenth-century self-presentation and Mouhot did not claim it. He said plainly that he was following where others had been.

Cowing, The Indianized States of Southeast Asia (1968) · Henri Mouhot · Public domain

Angkor Wat was never abandoned. It was maintained as a Theravada monastery continuously from the sixteenth century at the latest — probably from the fourteenth — with Buddhist images installed, monks resident, and a pilgrimage traffic. There are Japanese pilgrim inscriptions in it from the early seventeenth century, one of them accompanied by a floor plan drawn in Japan. Portuguese and Spanish missionaries described it in the 1580s and 1590s.

A sixteenth-century Khmer king had substantial restoration work done and had the unfinished bas-reliefs completed. That is a functioning building, not a ruin.

The other temples were a different matter. Angkor Thom, the Bayon and the outlying complexes were largely reclaimed by forest, which is what happens to brick and laterite in a monsoon climate when nobody is sweeping them.

What the French did

The French protectorate from 1863 produced the archaeology, the conservation, and a great deal else.

Wellcome Collection · Cowing, The Indianized States of Southeast Asia (1968) · John Thomson · CC BY 4.0

The École française d'Extrême-Orient cleared, surveyed, mapped and restored on a scale nobody else could have funded, and the scholarship — Coedès above all — established the chronology of the inscriptions and made a connected history of the Khmer empire possible for the first time. Everything anybody knows about Angkorian kings rests on that work.

It also came with the ordinary apparatus of colonial appropriation: casts and sculptures removed to Paris, a reconstructed Angkor pavilion at the colonial expositions, and an argument — advanced explicitly — that the Cambodians of the nineteenth century could not be the heirs of the people who had built this and therefore needed a guardian.

André Malraux was arrested in 1924 for taking sculptures off Banteay Srei, which is a fair summary of the period's range.

The flag

Cambodia is the only country in the world whose national flag has a building on it.

Angkor Wat has been on it, in one form or another, since 1863 — through the protectorate, independence, the Khmer Republic, Democratic Kampuchea, the Vietnamese-backed republic, and the restored kingdom. Every regime that has governed Cambodia in a hundred and sixty years, including the one that murdered a fifth of the population, has put the same twelfth-century temple on its flag.

The Khmer Rouge version was yellow on red. The current one is white on blue and red. The building does not change.

What is actually inherited

Not the state, which was gone by the fifteenth century, and not the religion of the temples, which changed while they were still standing.

The language, which is Khmer and is directly continuous. The Theravada Buddhism that replaced the temple cult and that Cambodia has practised for six hundred years. And the fact of the monuments themselves, which have been the single fixed point in every account Cambodians have given of who they are since the empire that made them stopped existing.

That is a strange inheritance and a real one. The buildings outlived the state by six centuries and became the argument for a nation the state would not have recognised.