The Site
The Sambor Prei Kuk Temple cluster is located in central Cambodia, about 80 Miles north of Phnom Penh, the capital. The temple cluster is located in a forested area in which 3 groups of temples are spread within an area of 3 miles and contain, all together, over 100 listed monuments.
The temple complex that makes up the modern archaeological site of Sambor Prei Kuk has been identified as the capital of the kingdom of Īśānapura founded by Īśānavarman (r. 616-635 ce ). The Chinese knew this region as Chenla and wrote about it as a unified state.[1] But rather than a unified state, Chenla was in fact a loosely tied together set of polities in constant flux. At times there would be less cohesion, and then under the leadership of particularly strong ruler, such as Īśānavarman I (r. 616-635 ce ) or Jayavarman I (r. 657-690 ce ), they would come together with the power radiating out of the center, only to separate out again as the power center either shifted or lost legitimacy.[2] Thus, we see here Tambiah’s pulsating galactic polities functioning in the manner of maṇḍala states.[3] With this political situation, the ritual theater of the state would play an important role in strengthening the tenuous, shifting bonds between these polities.
The territory under the control of Īśānavarman in the 7th c. ce extended out from the capital at Īśānapura. In the general Sambor region, ten temple complexes—and hundreds of individual temples—have been found in a twenty square kilometer area.[4] At the center of this square are three temple complexes that make up the ritual, political, and social capital of the region. These three complexes are referred to as Sambor Prei Kuk proper.
Each group—labeled the northern (N), central (C), and southern (S) groups for convenience—follows the same pattern: there is a main shrine in the center enclosed by two rectangular walls. This creates an inner courtyard and an outer courtyard. Thus, each group functions as its own maṇḍala ; one can readily see the squares embedded within each other. The northern group has fourteen structures (N1 – N14 ) distributed throughout the two courtyards. The southern group has ten structures (S1 – S10 ) all clustered within the inner courtyard. And the central group has three structures, the central sanctuary (C1 ) and two smaller shrines, which are not extant but are found in the archaeological record, within its walls.
The layout, architecture, and sculpture of the three groups of temples at Sambor Prei Kuk would serve the ritual theater of the state quite well. It is here that the vision of a cosmic maṇḍala would be instantiated in the land and ruler. There is much evidence that the religious élite practiced Śaivism, and thus at the supra-mundane level Śiva would be the deity at the center of cosmic maṇḍala , and this would be replicated at the state level with the king as the earthly counterform of Śiva who is the source of power throughout the realm. This is reflected in Īśānavarman’s very name: the Sanskrit prefix to his name, īśāna , has both political and religious implications as it means “a ruler, master, one of the older names of Śiva-Rudra . . . the sun as a form of Śiva.”[5] This connection between the ruler and Śiva is made explicit by O. W. Wolters,
The overlord’s close relationship with Siva (sic) meant that he participated in Siva’s divine authority . . . . He participated in sovereign attributes of cosmological proportions . . . . The chief’s prowess was now coterminous with the divine authority pervading the universe.[6]
The prefix īśāna is also applied to the city itself, Īśānapura, “The City of Śiva,” once again invoking Śiva as deity and Īśānavarman as his earthly counterform.
While many of the temples are in ruin and most of the iconography is now absent, there are sixteen extant inscriptions that have been found at Sambor Prei Kuk. With these clues, we can piece together a provisional picture of what the ritual world at Īśānapura/Sambor Prei Kuk entailed. A pilaster on the eastern gate of the outer wall of the Southern group of temples contains an inscription dated to the reign of Īśānavarman and dedicated to Śrī Prahasiteśvara—a form of Śiva found in Pā ṭ aliputra in the region of Magadha, India—which describes the images in the central sanctuary (S1 ). It reads:
Victory to Śrī Prahasiteśvara . . . (K. 440 v. 1)
. . .
These four images of Harihara, Ardhanārīśvara, Śiva [and . . .], this golden Liṅga, together with [an image of] the Four-faced [Brahm ā], this image of Śiva ……; this image of Sarasvatī, and this of Nṛtteśvara: all this has been installed by that king. He has further installed th is silver image of [Śiva's] Bull, which seems to be the body of Dharma in the K ṛ ta Age, [when it was still] undiminished. (K. 440 vv. 31-34)[7]
The central icon of the shrine is a golden Liṅga, the aniconic form of Śiva which received the standard worship by Śaivite initiates and the royal family. But in this shrine we see the aforementioned layered nature of Śaivism, and we can see in the central sanctuary S1 the pattern described by Sanderson,
There was a single Śiva at the heart of each foundation, generally embodied in a Liṅga, who received a version of the regular worship that initiates were required to perform for themselves. But the sites also enshrined (1) ancillary Śiva forms that had no role in the higher worship of the initiated, but evoked the mythological dimensions of the deity that are so central a feature of lay devotion, and (2) images of various other deities besides.[8]
Thus, the S1 shrine itself can be imagined as a self-contained maṇḍala with the powerful Liṅga at the center, then radiating outward lesser forms of Śiva (Harihara,[9] Ardhanārīśvara,[10] Nṛtteśvara,[11] and perhaps a standard two-armed single-faced Śiva[12] ), and finally other deities from the brahmanical pantheon (Brahmā and his consort Sarasvatī). As we move to away from the center of the shrine, the iconography becomes less Śaivite and more catholic in nature.
But Īśānavarman also explicitly detailed how the whole region he controlled fit into a Śaivite realm. At the west gate of the outer wall of the very same S group, there is an inscription (K. 441) which complements—or mirrors, as it resides on the opposite end of the east-west axis—the above inscription from the east gate. Rather than detailing the maṇḍala of the central sanctuary S1 , it inscribes Īśānavarman’s whole realm as a maṇḍala of Śiva. It reads:
I shall indicate the measure in height and circumference of the Liṅga on Liṅgadri and of the mountains in Liṅga pura. The Liṅga measures fifteen fathoms and a half [25 m][13] in height and fifty-eight fathoms [92 m] in circumference. From the base of the Liṅga on the ground of the plain to the mountain top, it measures two thousand seven hundred and twenty fathoms [4,350 m]. From the town named Giriśa to the mountain[top], it measures one thousand seven hundred and twenty-six [fathoms 2,760 m]. The three sacred waterfalls, attended by a legion of Sages, imitate the triple current of the Ganga at all times. The two dimensions of the esplanade at the top of this mountain are twenty-two fathoms [35 m] wide and thirty fathoms [48 m] long. (K. 441 vv. 1-6)[14]
While the east gate inscription (K. 440) functions to describe a shrine with Śiva at the center, the west gate inscription (K. 441) functions to integrate Īśānavarman’s mundane territory with Śiva’s cosmic realm.[15]
[1] Understanding the nature of Southeast Asian polities as maṇḍalas helps make sense of the various names given to the region. It was originally identified as Funan by early Chinese records, but soon after the 7th century ce , and as the center of control in the region moved inland, it became known as Chenla. Even later, it was referred to as Kumbujadeśa. But Xuanzang, the famous Chinese monk who detailed much of the geography of South and Southeast Asia in his court records of the 7th c. ce , calls it by none of these names, but rather identifies it with its ruler Īśānavarman and calls it Īśānapura (see Coedès, Les États Hindouisés d’Indochine et d’Indonésie , p. 87 ). Clearly, the region is loosely held together and not a fixed “state.” For an overview of the region known as Funan, Chenla, Kumbujadeśa, and Īśānapura, see Dougald O’Reilly, Early Civilizations of Southeast Asia (Lanham: AltaMira Press, 2007), pp. 91-126 . For a detailed discussion of the debate surrounding the name and location of Īśānapura see Vickery, Pre-Angkor Cambodia: The 7th-8th Centuries , p. 410, note 61 .
[2] Vickery’s reconstruction of Īśānavarman’s reign fits quite well with the maṇḍala model. For example, Vickery writes, “[t]he pattern of most of Īśānavarman’s inscriptions, then, does not demonstrate that he ruled ‘Cambodia,’ but that his authority was strong in Kompong Thom and Prey Veng (Ba Phnom), and was decreasingly exerted toward the South where local élites merely evoked his suzerainty while maintaining their own local authority . . . Outside his core kingdom in Kompong Thom the records suggest rather autonomous local chiefs sometimes voluntarily acknowledging some kind of super-ordinate hierarchy, but not subject to direct rule by the suzerain.” Ibid., p. 337 .
[3] See Brown, The Dvāravatī Wheels of the Law , pp. 13-18 for a detailed description of such waxing and waning of polities spanning the 7th and 8th centuries in Cambodia.
[4] For a general overview of Sambor Prei Kuk and its surrounding region, see Dumarçay and Royére, Cambodian Architecture , p. 39-44 and figs. 34-40 .
[5] Monier Monier-Williams, Sanskrit-English Dictionary (Delhi: Motilal Banarsidass, 1984), p. 171 . Īśāna can also be a name of Viṣṇu, but as the balance of this article will show, the preponderance of Śaivite iconography and inscriptional references clearly points to its identification as a Śaivite name.
[6] Wolters, History, Culture, and Region in Southeast Asian Perspectives , p. 11.
[7] For a full description of the inscription, see Coedès, Inscriptions du Cambodge , vol. IV, pp. 3-11 . The translation used here is by Sanderson, “The Śaiva Religion Among the Khmers,” p. 436 and n. 314 . It should be noted that Vickery, Pre-Angkor Cambodia: The 7th-8th Centuries , only mentions this inscription once in the context of placing the inscriptions in their geographical locations. This is a glaring omission as it is a lengthy inscription found in situ . However, by ignoring this inscription, and many others with religious content, he is able to make his argument for the primarily “non-religious” nature of the pre-Angkor inscriptional corpus as well as contend that Indian forms of Śaivism had no real impact on the region.
[8] Sanderson, “The Śaiva Religion Among the Khmers,” p. 436.
[9] Harihara or Śaṅkaranārāyaṇa is an image in which the deity is half Śiva-half Viṣṇu. See Helen Ibbitson Jessup, Art & Architecture of Cambodia (New York: Thames & Hudson, 2004), p. 42 fig. 38 , for an example of a 7th c. Khmer sandstone image of Harihara from shrine N10 at Sambor Prei Kuk. For other examples, see Ibid., p. 32 fig. 32 , a full size 7th c. sandstone image from Angkor Borei; Stefano Vecchia, The Khmers: History and Treasures of an Ancient Civilization (New York: White Star Publishers, 2007), p. 32 fig. 32 , a 7th c. head from Prasat Phnom Da; and Donatella Mazzeo and Chiara Silvi Antonini, Ancient Cambodia (New York: Grosset & Dunlap, 1978), pp. 27-28 , a full size 7th c. image from Ashram Maharosei.
[10] Ardhanārīśvara is an image which is half-Śiva-half Umā, Śiva’s female consort. While there is no extant Ardhanārīśvara from the 7th c., we do have a bust of Umā from the central sanctuary of the northern group (N1 ) at Sambor Prei Kuk. See Ibid., p. 31 .
[11] Nṛtteśvara is a dancing Rudra (Śiva). I know of no extant 7th c. images of Nṛtteśvara.
[12] See Jessup, Art & Architecture of Cambodia , p. 53 fig. 50 , for a 7th c. figure from Kompang Cham Kau.
[13] One fathom (Sanskrit vyāma ), is about 1.6 m, see Coedès, Inscriptions du Cambodge , vol. IV p. 15 .
[14] Ibid., vol. IV pp. 14-17 . The text here is my translation of Coedès’ French translation of the Sanskrit. Coedés finds the measurements somewhat plausible. The distances from Giriśa to the Liṅga are not the height of the mountain, but rather the length of the path from the town to the mountain-top. As Coedés argues, “Quant à la hauteur des montagnes, il ne s’agit évidemment pas de leur altitude au-dessus du niveau de la mer ou du pays environnant, que les anceins Khmèrs povaient difficilment èvaluer, mais de la longueur du chemin qui menait de la basee au sommet. Or, 4 km 350 [4,350 m] et 2 km 760 [2,760 m] sont des distances parfaitement admissibles pour des hauteurs de 500 à 1.000 ou 1.500 mètres” Ibid., vol. IV p. 15 . Coedés suggests the mountains are most likely to the north and lists the following mountains and their heights: Phau Kao (1397 m), Phou Bassac (1403 m), Phou Douang (1270 m), Phou Phaphine (875 m), and Phou Malang (1286 m).
[15] For a comprehensive study of how both pre-Angkor and Angkor rulers mapped Indian Śaiva geography onto the Southeast Asian landscape, see Sanderson, “The Śaiva Religion Among the Khmers,” pp. 403-421 . The most interesting part of this description is the identification of “self-born” Liṅgas by the Khmers, that is natural, not man-made, Liṅgas that spring from the very land itself, see Ibid., p. 411, n. 242 .