The Cambodian expansion

The Cambodian expansion

Dharmacakra, from U Thong. Detail. U Thong National Museum

The Cambodian expansion

Once a capital was established at Angkor around A. D. 900, the Khmers consolidated the monarchical, social, religious, and architectural institutions that defined Cambodian life for three hundred years. The political power of the central court rose and fell, and political control over the affected parts of Thailand fluctuated, but everywhere it is the cultural life of Cambodia that allows historians to give shape to developments in Siam. Khmer cultural influence was followed by a development that could hardly have been fore- seen—the swift collapse of classical Khmer civilization after about 1200. It is almost as if the domination of central Siam was a victory that could only be won at the expense of the defeat of the traditional institutions. To these two great events—the Khmer expansion and subsequent cultural transformation—must, from the perspective of 1300, be added a third, namely the arrival in Thailand of a new ethnic group, the Thais, who by 1300 stood in a position of supremacy.

The period between 900 and 1300 brought also the establishment of enduring iconic and architectural forms. One is the Buddha image in the earth-touching gesture, today generally called in Thai Maravi- jaya, victory over Mara. Another is the type of tower known as the prang—initially a sanctuary tower, later a tower with four false doors. A third is the crowned Buddha image. To say that these elements were a Khmer gift to Thailand would not be exactly correct. True, all three are first found together in a Khmer temple, the Buddhist temple of Phimai, dating from the years around 1100. But the Tantric Buddhist elements found at Phimai are from the Khmer point of view intrusive. The language of art and architecture that gave substance to the crowned and earth-touching Buddhas merely happened to be Khmer. Even the sanctuary tower need not have taken a Khmer form, though under the circumstances it did, creating a enduring legacy. The word prang first appears as pran, meaning “high tower,” in an inscription of 639 (K. 79), and so it is probably of
Khmer origin, but by the tenth century it was assimilated to Sanskrit prangana, “courtyard.”
The establishment of these iconographic and architectural types might well have occurred, therefore, without any Cambodian expansion at all. Meanwhile, many of the cultural elements that can be considered classical Khmer were rejected, either swiftly or over a period of years. Stone carving never took hold, except for a period at Ayutthaya. The carved stone lintel never entered the local repertory. Stylistic concerns evident in Khmer sculpture—the interplay of volumes, the sharp textural distinctions, and the role allotted to horizontal elements—were never fully embraced. Indeed, they were challenged and made obsolete. The iconographic types most strongly associated with the official Khmer Buddhism of the late twelfth cen- tury—the naga-protected Buddha and the Mahayana divinities Avalo- kitesvara and Prajnaparamita—were rejected with varying speed in the course of the thirteenth century. Khmer Hinduism had no lasting popular impact.
This pattern of rejection was certainly not total. Other phenomena could be explored: memorial images, images sacred to particular localities, court ceremony, funerary monuments. Not all such features need have had Cambodian origins, however. The extent of the debt is hard to measure, due to a lack of knowledge about both Dvaravatl and Tai traditions. It is reasonably certain that there were local continuities. But no full sequence of locally made objects—in brick, stucco, or bronze—has yet been identified as belonging to the tenth, eleventh. Yet it is not possible to grasp in any concreteness the societies doing the rejection, or to assess their relationship to Dvaravatl traditions.

The Northeastern Provinces
In 868 a Khmer speaker named Ansadeva established a golden linga somewhere in or near Muang Sema, in Nakhon Ratchasima province, according to an inscription that describes the place as inside Kambudesa (Cambodia) and abandoned (K. 400). This was recorded in an archaic script on the reverse side of a stone slab commemorating the much earlier donation by the lord of Sri Canasa. The earlier inscription and the uncarved boundary stones that might date from the same time were described in the last chapter (pp. 99-100 and fig. 16). It is symptomatic of the mix of activities that were taking place in these border regions in the ninth century—and later as well—that a stone slab discovered in the same area has upon it an image of the meditating Buddha, above him, within a niche of Cham style, the representation in low relief of a stupa. This stele cannot date from long before or after 868. It was a time, therefore, when local scribes were writing Sanskrit in an old-fashioned way; when local Buddhism was still being followed; when Cham influence had penetrated the region; when Saivaism was spreading; and when, meanwhile, the territory of Cambodia was growing. The kingdom of Sri Canasa was itself not dead in 868, but where it lay and what territories it controlled are not known. It lasted for at least another seventy years, for a stone inscription of 937, giving the name of the adhipati of Sri Canasa, was discovered in Ayutthaya (K. 949).
In the decades following 868, nevertheless, evidence for Khmer activity in the region is abundant. At the temple site of Phanom Wan, little more than fifty kilometers to the east of Muang Sema, a lintel of late ninth-century type was discovered; it can be associated with a door-jamb inscription (K. 1065) of Yasovarman (r. 889-900) that speaks as well of the authority of the king’s predecessor Indravarman I (r. 877-89). Sometime later, at a site ten kilometers west of Phanom Wan, a short inscription (K. 396) was set up that includes the posthumous name of Harsavarman I (r. 912-22). (A bronze Prajnaparamita said to come from Prachinburi province may indicate that the Prakhon Chai school of bronze-casting remained vigorous until the early tenth century.) On nearby Phanom Rung, two brick sanctuaries were built around the same time. An inscription that may date from the Jayavarman V period (ca. 968-1000) refers to the site as Vnam Run (“the broad [i.e., flattop] mountain”) and speaks of a devasthana, while another inscription (K. 1068) mentions gifts to the Kamraten Jagat Vnam Run—the Lord (principal image?) Phanom Rung. Around the middle of the tenth century, finally, a splendid temple, Prasat Mfiang Kh*k, was established a few kilometers east of Muang Sema itself. The lintels from this temple may be somewhat less refined than others made nearer the capital in the same period, but the central figures—Indra, Durga Mahisasuramardini, the dwarf incarnation of Visnu—were carved in high relief with astounding vigor.

Figure 19. Eleventh-century Khmer temples ground plans. (a) Muang Tam. (b) Kamphaeng Yai. (c) Phanom Wan.

Figure 19. Eleventh-century Khmer temples ground plans. (a) Muang Tam. (b) Kamphaeng Yai. (c) Phanom Wan.

It may be correct to consider Muang Kh£k (or another nearby tenth-century monument, Non Ku) provincial in style, but this is primarily because it stands at a certain distance from a cosmopolitan structure of the period, not because its characteristics can be described in terms of an older tradition, where local Buddhist iconographic traditions were incorporated, in a way not possible in the earlier tenth- and eleventh-century Hindu shrines. But within this broad framework there were variations, as the introduced Cambodian styles give birth to their own local traditions. The Prasat Myang Kh£k lintels have a frieze at the top filled with a row of seated hermits having beards, crossed ankles, and hands in adoration. In the course of the eleventh century, the frieze of hermits seems to have been somewhat more significant element in lintels at various sites in the Northeast than in the lintels at temples south of the Dong Raek range.
This corridor, as it developed in the second half of the tenth century and in the course of the eleventh, apparently became an increasingly integral part of the kingdom of Cambodia; meanwhile the area around Muang Sema at the western end of the Mun River watershed decreased in importance. Traveling toward Phimai along this highway meant passing by important temples and population centers. At the modern Cambodian border, where a pass through the hills leads to the Cambodian plain, is the eleventh-century Prasat Ta Myan Thom; an inscription (K. 376) attests to tenth-century activity in the area. In Ban Kruat district of Buriram province, west of Ta Myan Thom, lie another temple, Prasat Bai B£k, and an important kiln site.
Cambodian brown-glazed wares, the dominant type, are unusual among the world’s ceramics for their silhouettes, which frequently have stronger parallels in metalwork and stone architectural elements than in earlier pottery. The introduction of pale green celadon and chocolate stoneware glazes—probably in the late ninth or early tenth century—is no doubt due to contact with Chinese potters, perhaps of Kuangtung province. Chinese shapes can be seen in bowls and spouted ewers, but they are the exception. Production of pale green wares ceased after the eleventh century, but brown-glazed wares were made at least into the thirteenth. Thirty kilometers bring the traveler to an area with three important complexes: Phanom Rung, Plai Bat Hill, and Mfiang Tam. Tenth-century activity at Phanom Rung has already been mentioned; at some point in the first half of the eleventh century the site was enriched by a sandstone sanctuary. Prasat Plai Bat (2) may date from the early eleventh century. After another fifty kilometers the traveler will arrive in the area of Muang Ban Fai, an ancient site where Prakhon Chai-type bronzes were discovered (p. 107 above). Phimai stands yet another thirty kilometers away, but about half that distance brings the traveler to the environs of Muang Phlap Phla, a town perhaps as old as Muang Fai but one that remained important into the period of concern, for an aerial photograph shows a rectangular, Khmer-type city superimposed upon an irregularly oval one characteristic of the Dvaravatl period.
After a spate of mid-tenth-century activity around the time of Rajendravarman (r. 944-68), the amount of construction in the Northeast seems to have dropped off. In the last decades of the tenth century and first part of the eleventh, there was activity at Angkor, but the precise sequence of monuments is not well understood. Along the southern edge of the Dong Raek mountains, near Chong Tako, a pass further west than the one by Prasat Ta Mfian Thom, a Saiva temple was established in 1007 (Phnom Sanke Kon or Khao Lon, Aranyaprathet district, Prachinburi). The inscription states that Suryavarman had then been ruling since 1002 and records the establishment of a linga and of images of Sambhu (Siva) and Devi (K. 232). (Suryavarman’s power was not consolidated until 1011.)
Work at Prasat Myang Tam (fig. 19a) can be divided into two phases that may have followed closely one after the other, or perhaps overlapped—first the five brick sanctuaries upon a platform (pl. 32), together with two libraries, and second the enclosing system of an inner gallery, four ponds, and an outer walled enclosure with four gopuras, all in sandstone (pl. 33A). The lintels on the brick shrines might be contemporary to the Khao Lon lintel of 1007—the central kala masks have some features in common—but otherwise they belong to different families. The lintel at the northeastern shrine (pl. 32) shows Siva seated in sukhasana upon the bull Nandi, his right arm holding a trident, his left around his spouse Uma. The lateral branches are horizontal, and across the top is a frieze of seated hermits, somewhat narrower than that at Prasat Myang Kh£k, but of the same type.
The lintels of the second phase at Prasat Myang Tam, that of the gallery and gopura system (pl. 33A), are in character similar to those at a number of sites, especially Preah Vihear, where most of the structures must date from between 1018 and 1049, the time- span of the inscriptions at the complex (K. 380). The shrine lintel (pl. 32) reveals, in comparison, fewer striations and contains more non-vegetal elements. The sculptor of the gopura lintel (pl. 33A) was more aware of cosmopolitan currents and gave to the pendant foliage a steady rhythm of small units and a hierarchy of larger ones, which both keep the eye busy and provide it with an overall structural order. The shrine lintel lacks this steady rhythm and easily perceived hierarchy of parts. As can be seen in pl. 33A, other features connect the Mfiang Tam galleries and gopuras to Preah Vihear as well, most particularly the design of the pediments, with a plain ground lying between the inner foliate fan and the pediment frame. Another common characteristic is the bare-headed naga, seen as terminant of the pediment frame in pl. 33A and a prominent feature of the railing around the Mfiang Tam ponds. All this suggests that when the local political powers established the brick shrines they were not able to draw on craftsmen who were up to the standards of the royal workshop; a little later, in the 1020s or 1030s, the time of the sandstone gallery and gopuras, they were. Within the gallery and gopuras, there are two main types of lintel, in quarters (pl. 33A) and with horizontal branches, and two main types of pilaster, with stalks either winged (pl. 33A) or fan-outlined. The distribution of the types is orderly, suggesting that the layout was planned at a single moment in time. A third pilaster type, with spirals, is confined to the interior of the outer eastern gopura.
Prasat Kamph*ng Yai (Uthumphon Phisai district, Sisaket) lies in another region altogether, to the east, north of Preah Vihear. The overall plan (fig. 19b) is similar to that at Mfiang Tam, with sanctuaries upon a common base, northern and southern libraries entered at the west, and enclosing galleries with balustraded windows, made of two shades of sandstone. The shrines and supporting base are of laterite, brick, and sandstone. An inscription in the eastern gopura (K. 374) records the dedication of land to the temple in 1042; though neither the content of the inscription nor physical evidence makes it possible to determine how much of the temple was completed before 1042 and how much after, a period around the 1040s fits in with other evidence. On the striking inner lintel from the main sanctuary, Indra sits on a profiled Eravana, and a frieze of hamsas stretches across the top (in the position of the previously encountered hermits). Where the branch is divided into quarters, it is surmounted on each side by a Garuda—a feature not seen at Mfiang Tam but paralleled at Sdok Kak Thom (Ta Phraya district, Prachinburi) in 1052 or thereabouts. In the northern and southern libraries, the inner and outer lintels differ in type, the inner lintels being of a pictorial type that developed at some point in the course of the eleventh century and not seen at Myang Tam. As can be seen in pl. 33B, the outer lintel of the southern library, divided into quarters, has at the center a rarely depicted Gaja Laksmi while on the inner pictorial lintel appear Uma Mahesvara. The fourth sanctuary, that of the southwest, may have held images of the same pair, judging by the width of the platform.
It is less easy to draw conclusions about the development of freestanding stone sculpture, although eventually petrological analysis may make it possible to attribute works now scattered around the world to northeastern Thailand (and to sort out certain items of recent manufacture). A pair of stone guardian figures excavated at Myang Tam belongs to the same period as the galleries and gopura and falls within the mainstream of Khmer sculpture. So does a female divinity uncovered at Phanom Rung, a work very likely contemporary with the building of the brick sanctuary that stands south of the main shrine. Another major work (pl. 34), in bronze, has been identified as a guardian, perhaps Nandikesvara, chief of Siva’s troops, but it may simply represent Siva. This sculpture, excavated at Kamph^ng Yai, attains the full mastery of the greatest cosmopolitan Khmer sculpture, the controlled sinuous curves of facial features and of the details of the jewelry both contributing to and softening an aloof presence. The character of the modeling and details of the jewelry resemble that of the giant head of the reclining Visnu recovered at Angkor, which is probably a work of the Udayadityavarman period (1050-66), when once again there was major construction at Angkor. The Kamph*ng Yai bronze was most probably made shortly before Udayadityavarman’s reign. The alloy (96% copper, 4% tin) indicates a stronger connection with Angkorian casting practices than with those that have been attributed to the Phimai region, where higher proportions of lead and tin seem to have been the rule. A large circle of bronzes in western collections, some of them depicting adorants or worshipers, can be attached to the same school, which was evidently productive in the years following the end of Udayadityavarman’s reign in 1066.
Yet another temple with an enclosing gallery system of two- colored sandstone is Prasat Hin Phanom Wan at the far end of the northwest highway, where the earliest activity dates from the late ninth century (fig. 19c). The tower-sanctuary of the eleventh-century main temple is reached through a forechamber and neck, much as at Preah Vihear or, later, at Phimai and Phanom Rung. In 1055, an inscription (K. 393) was cut at the sanctuary door, and an inscription of 1082 (K. 391) provides evidence of continued activity at the site.
This inscription mentions King Jayavarman (VI) and thus is one of the pieces of evidence relevant to the troubled political history of the period. Udayadityavarman, who built the Baphuon at Angkor and reigned from 1050 to 1066, was succeeded by his brother (r. 1066-80), whose rule seems to have been weak, for he was followed not by a member of his family, but by a king who established a new lineage, known as the Mahldharapura dynasty. In 1065, in the reign of Udayadityavarman, there had been a revolt—Kamvau’s revolt. The population centers along the Angkor-Phimai road were somehow involved; in 1067, Harsavarman’s nephew restored a linga that had been smashed by Kamvau’s men at Prasat Preah Khset, a site halfway between Angkor and Ta Mfian Thom at the modern border (K. 237). One possibility is that the Mahldharapura dynasty had its origins along the Angkor-Phimai route, perhaps even in Phimai itself. It seems more likely, however, that the family’s home was further east, around Koh Ker. At the temple of Phnom Sandak, northwest of Koh Ker, the inscription most elaborate in its praise of monarchs of the dynasty was cut in 1110 (K. 191), and in 1119 the official who had been instrumental in the family’s rise to power left an inscription at the same temple (K. 194). Furthermore, if it is asked what the royal workshops were doing following the completion of the Baphuon, a good response is that they were building Prasat Khna Sen Kev, a temple about fifteen kilometers southeast of Koh Ker—in the home territories, it can be surmised, of the most powerful family in the kingdom. If the Mahldharapura dynasty monarchs did originally come from this area, then the local officials or chieftains along the Angkor-Phimai road must be understood as allies and supporters of the Mahldharapura monarchs—and perhaps for a period more powerful than either the first Mahldharapura king, Jayavarman VI (r. 1080-1107), or the second, Dharanlndravarman (r. 1107-13). Local officials apparently had the power to recruit the finest craftsmen in the kingdom, thus resulting in the construction of the great temple of Phimai in the decades following the building of Prasat Khna Sen Kev. In the far northeast, a similar situation may have prevailed, for extensive work at Vat Phu was carried out in the same Jayavarman VI period.

North of the Mun
From the late tenth century onward it is possible to point to four different sorts of development north of the Mun. First is the gradual northward encroachment of both Cambodian political power and art styles. Second is the adaption by the older boundary-stone culture of Khmer stylistic elements, accompanied, probably, by geographical movement to more remote sites—though little can be surmised about the political situation. Thirdly, there seem to have been certain long-range Khmer thrusts—up the Mekong River, for instance. And finally there is the presence at the site of That Phanom (Nakhon Phanom province) of a culture distinct from that of both the boundary- stone and the Angkorian traditions.
It is possible to observe the spread of both Cambodian styles and political power up the tributaries that flow southward into the Mun— for instance, in the area around Prasat Ku Krad§n (Kaset Wisai district, Roi Et). An entirely pictorial lintel there, probably of the second half of the eleventh century, bears a scene from the Ramayana. Meanwhile, another thirty kilometers or so to the east an inscription at Ku Aram (K. 373), probably of the eleventh century, lists the names of Cambodian officials. North and west, this picture of imperial integration continues. A trip of about thirty kilometers up these tributaries brings one to the vicinity of Muang Champasi in branch district Na Dun, Maha Sarakham, a site that flourished in Dvaravati times (see fig. 17d), and where Mon was spoken. Near or at Champasi have been found fine examples of eleventh-century Khmer sculpture, including a naga- protected Buddha. Later—at the end of the twelfth century— laterite shrines were built here.
The people of Champasi had been builders of stupas and modelers of stuccos, and at an unknown moment this they apparently ceased to be. Such is equally the case at Fa Daet, in the nuclear Chi region, about eighty-five kilometers north of Ku Krad§n. There are no Khmer sites near Fa Daet; perhaps the population was displaced. There are reasons to believe that some of the sima carvers of Fa Daet were forced to move into outlying regions. On stylistic grounds, sima of various types appear to postdate pl. 27 (the Fa Daet sima illustrating the Sarabhanga-jataka), and some of these can be considered developments of the Fa Daet type. Sometimes the general profile was kept, but the proportions were much attenuated, and foliage appeared at the base. This type can be found on the upper reaches of the Chi, at Muang Chai Wan in Manchakhiri district. Much further north, in Ban Phu district of Udon, Khmer-influenced niches with figures replaced the foliage in a sima of similar propor- tions. Simas of square section are another type—one paralled in the Khmer repertory. Examples have been found at the square town of Nong Han in Udon province, and at the site of Ban Huai Hin Tot, about 60 kilometers northeast of Fa Daet.
These examples come from areas that lie a certain distance from Fa Daet, where such sima have not been found. The notion that the sima-carvers of Fa Daet were driven away and became re-established elsewhere is arguable on the basis of pl. 35, a detail of one of the sima from Ban Huai Hin Tot. Piriya Krairiksh identified the dancing figure as the naga princess who sings for a husband on Black Mountain in the Vidhurapandita-jataka. Many of the elements are similar to those in pl. 27—the height of the relief, the platform tilted to create space, the interest in conveying movement, the cloth (whether banners or scarves) in low relief. At the same time, the development is an interesting one: the vivacity and lightness of the figure is conveyed in part by her isolation and by the interplay of rhythm and scale between the two legs, two arms, and two ends of the scarf.
According to Dr. Piriya, this is the second in a sequence of scenes running counter-clockwise, the first being an earlier episode from the Jataka, the third being a seated figure identified as the Bodhisattva Maitreya (the fourth side having foliage). Inscribed beside this seated figure is a Khmer or Khmer-sounding title (-kammraten) followed on a second line by a personal name that has not been entirely read (K. 510). The figure might, therefore, be the donor rather than Maitreya. Beneath these figures in relief—buried in the earth in the view of pl. 35—are bisymmetrical scrolls of foliage, having the organization of a Khmer lintel of the tenth or eleventh century, by which the carver was probably influenced. The sima, therefore, seems to be a positive response to a Khmer stylistic challenge of (probably) the late tenth or the eleventh century. It is not possible to say whether the title had been bestowed upon the presumed patron by Cambodian rulers or was one used in an independent or semi-independent political entity.
The upper reaches of the Chi take one not just to the site of Chai Wan but toward the hills that mark the western edge of the northeastern region. Here is the reclining Buddha of Phu Wiang Mountain (pl. 28). Perhaps there was a kind of culture of the hills, extending both south and north, in which natural rock formations had a sacred value. At Phu Phra in Chaiyaphum a row of seven seated Buddha images was carved in a rock shelter, next to a Buddha image carved in reverse Maravijaya. The striated hair, which comes to points on the temples, suggests the work of an untrained sculptor, following a vaguely Khmer model, perhaps as early as 1000.
The intermediary site is Wat Ph§ Ta near Phra Phutthabat Bua Bok, on the Phu Phan range, close to the prehistoric paintings of fig. 5. Here in the rock is carved a Buddha, seated in meditation in a well-carved triangular niche with lozenges on the frame and scrolled terminants, such as can be found at Banteay Srei. On a smaller scale are figures of the standing Buddha. One can be seen in pl. 36A. The pose, with left arm at the side and right hand performing a gesture, is one that became popular in late Dvaravatl times. It appears at Vang Sang and had entered the Khmer repertory somewhat earlier, in about the middle of the tenth century (pl. 36B is an example). Other conventions—the raised belt, the pleated undergarment visible between the legs—are late Dvaravati features as well. The curve of the left arm provides a stylistic link to the Dvaravati reclining Buddha images (e.g., pl. 28). A date in the second half of the tenth century, before rather than after Vang Sang, is plausible. This amalgam of Dvaravatl and Khmer styles also characterizes the tall sima mentioned above, with niches at the foot, found not far away in Ban Phu district. If Mon speakers were responsible, as seems possible, then perhaps they were long-established inhabitants, who at this point in time were able to acquire the services of a stone carver who knew Khmer styles well.
Phra Phutthabat Bua Bok is far enough north that it is necessary to take into account the Mekong River as a route of communication with the political and artistic centers of Camboda. Farther east in Udon province, as already mentioned, there is the square town of Nong Han. This Nong Han is the legendary kingdom Nong Han Noi, the “lesser Nong Han” (the other Nong Han—Nong Han Luang—is a lake on the edge of Sakon Nakhon). The principal monument in Sakon Nakhon, the Lao-style Phra That Ch&ng Chum, was originally a Khmer-style sanctuary, a door jamb of which preserves an eleventh-century inscription in Khmer, albeit with idiosyncratic local features (K. 369). Outside the town stands Phra That Narai Cheng Weng, a Khmer shrine having a lintel and a pediment stylistically akin to others carved much further south in the second half of the eleventh century. There is nothing in the physical evidence that need contradict a local legend: that the ruler whose queen established Phra That Narai Cheng Weng was the great-grandson of the king of Indapat, or Angkor, and the grandson of the city founder. Sometime later, after seven years of drought, the populace migrated to Cambodia.
The Klam flows from Lake Nong Han to the Mekong River at That Phanom. The brick tower of Phra That Phanom, which collapsed in 1975 and was soon rebuilt, is like no other monument in Thailand (pl. 38A). The original shrine, with its carved brick decor, is older than the Nong Han structures, which are in no apparent way local developments of it. The study of brick construction methods led Anuvit Charernsupkul to group Phra That Phanom with brick monuments that fall between Myang Kh£k and Myang Tam, belonging to a style vaguely Pre Rup and falling in the middle or second half of the tenth century. The carved bricks, comprising vertical panels that lie between pilasters (and engaged colonnettes) and the corners, on the other hand, have connections with the art of Champa—as found at Miso’n A-10 (perhaps early tenth century) or at still older monuments. A section of one of these panels appears in pl. 38A, on the right-hand side of the eastern or front face, flanking the original sanctuary entrance. The general character of the foliage, and especially its bilateral symmetry, can be matched in tenth-century Champa, as can the horse in flying gallop, a borrowing from China. The style of the foliage is not uniform on all four faces. On the north side, for instance, there is no bilateral symmetry and there are numerous pointed leaves, somewhat like those found on Northern Celadon ware of the early Sung dynasty. Yet the style of the elephants, horses, and figures is approximately the same.
In each one of the panels appear a man on a horse and a man on an elephant, the elephant and the horse taking top position in alternate panels. The animals and riders move toward the eastern entrance, the end of the two lines being the southwest corner, for the animals of the western side are moving in a clockwise direction on both panels. At the end of what might be considered the longer line, on the southern panel on the western side, the horseback rider, his hair tied in a bun at the back of his head, is accompanied not only by a parasol holder (as in pl. 38A) but by a figure carrying what has been identified as a fire box of the sacred fire. The legendary history of Phra That Phanom, a compilation which in its original form seems to date from the time of the installation of a Buddha relic and restoration in the late seventeenth century, identifies the carved brick figures as the ancient kings—the same kings who appear in the Sakon Nakhon legends mentioned above, from which the compilers of the Phra That Phanom history probably borrowed. The intuition—that the figures on elephants and horses are real people, not gods—feels correct.
Phra That Phanom may be considered a Saiva shrine of the tenth century. Cham elements of various sorts have been noted in different parts of Thailand. Phra That Phanom brings additional substance to the theme of Cham influence, a theme, however, hard to trace in the written evidence. The tenth-century That Phanom was built over an older brick monument, and so there must have been local cultural continuities. One possibility is that here was a commercial center brought into the Cham orbit through a marriage alliance. But with the Vietnamese raids on Champa and the Cambodian expansion, there was little chance for Phra That Phanom to bequeath a legacy of later Cham-style structures.
The local sculptors were capable of what seems to be creative response; of replacing the old stylistic language with a crude adoption of Khmer-like forms (Phu Phra, Vang Sang); and of what seems to be skillful adoption of Khmer motifs mixed with more traditional elements (Wat Ph§ Ta, pl. 36A). In a dynamic situation, the developments cannot be expected to be uniform.

Central Siam: General Considerations
It is the site of Khmer-type temples but not of sandstone monuments emulating those of the capital, as were the temples south of the Mun River. Nevertheless, in the far north, the Mon settlements of the kingdom of Haripunjaya (to be discussed in the following chapter) would represent a prolongation of Dvaravati civilization, touched but not overwhelmed by Angkor. In between these two poles (of Haripunjaya and Lopburi), there is a patchwork of varying developments, which intensive archaeology may or may not ever be able to transform into a high-resolution panorama. Even at a site from which there is a good deal of evidence.
It might even seem that speculation on the basis of floating objects is more fruitful. Perhaps, therefore, the absence of archaeological evidence, of inscriptions, and of adequate study of the material that survives has meant that the threads of continuity remain largely undetected or undefined. Take, for an example, the head of a bronze Buddha standing in double vitarka-mudra (pl. 37). The presence of small curls beside the forehead and on the temple is evidence of a connection with the arrangement in the Cham art of Dong-du’o’ng, an arrangement also reflected at Vang Sang, Laos (p. 131), in tenth-century Khmer art, and in a head of the Buddha—probably tenth-century in date— from Nakhon Ratchasima province. The fact that the incised lines at the top and bottom of the outer edges of the eyes do not join suggests the influence of tenth- or eleventh-century Khmer sculpture. Here, therefore, is a Mon sculptor (presumably) responding to the fashions of the times in a creative way. The conclusions one can draw from such an example are few in number, however, when it is not known where the bronze was made. The softly modeled eyebrows recall those seen in a late-Dvaravatl terracotta head (pl. 29A). Without a context, it is impossible to say what the next generation of sculptors was doing: whether they, too, were updating Dvaravatl traditions, or whether they were abandoning them.
This bronze Buddha image has Cham connections, and it is thus representative of an important aspect of the art produced in the period around the tenth century, seen in Cham-connected works ranging from That Phanom in the Northeast to Khfiha Sawan Cave in the South (pp. 133 and 114). In the later Dvaravatl period there were also elements that suggest a connection with Cambodian art (in stucco decor and in the design of a mask; above, pp. 115 and 97), but a connection that appears to have no apparent relationship to political expansion. The directions taken by Mon art in the tenth and eleventh centuries may not, however, merely be a matter of late Dvaravati tendencies, characterized by Cham and Khmer elements, finally succumbing to severe pressures from Cambodia. The reason is that in neighboring Burma the eleventh century brought the flowering of a Buddhist and largely Mon civilization, and this civilization was dependent in many ways upon that of Pala India. Pala- type votive tablets have been found in Siam, and at least some of them must have been fashioned or imported in the tenth and eleventh centuries. If this bronze was made in central Siam in the eleventh century, it could be used to demonstrate that Mon Buddhist art took a radically new turn in Siam, one that paralleled the important developments at Pagan in Burma. Here would also be evidence that many of the stylistic and iconographic features of fourteenth-century Sukhothai art had long-standing local roots. Once again, however, the absence of a context for the bronze image of pl. 39A, which was deposited at Wat Ratchaburana in Ayutthaya in 1424, makes such assertions entirely speculative.
Another aspect of Khmer expansion is the Khmer appropriation of iconic types—types which would then henceforth have a Khmer character to them. A key Khmer inscription and a key image both come from the same site—the brick sanctuary of Beng Vien (Siem Reap province)—and may date from the same time—in or shortly after 946. The inscription (K. 872) claims that Rajendravarman (r. 94468) won victories over Champa and Ramanya, that is, Monland. There are earlier mention of Mons in Khmer epigraphy, but they appear as slaves. Although it is not know what territories “Ramanya” consisted of in the tenth century, Rajendravarman’s victory was probably a significant one. The foundation, furthermore, was Buddhist, and the inscription provides the earliest evidence for the worship of a triad that much later—during the reign of Jayavarman VII— became the primary focus in a state cult. This triad consisted of the Buddha, Lokesvara (i.e., Avalokitesvara), and Prajnaparamita, who is called the begetter of the series of Jinas. The notion that the Perfection of Wisdom is the mother of Buddhahood can be found in the very earliest Prajnaparamita texts. Lokesvara, by implication, is the father of Buddhahood.
In Dvaravatl art this icono- graphic type is characteristic of the later period. It is the type seen in the Wat Ph§ Ta sculptures (pl. 36A), for which a date in the second half of the tenth century has been proposed, and at Vang Sang (A. D. 1006). In pl. 36B is illustrated yet another floating bronze (like pl. 39A, it was deposited at Wat Ratchabfirana in Ayutthaya). The way the robe falls over the left wrist is much like that in the Prasat Beng Vien image; the belt can be compared to that on the Buddha in the Wat Pho Ta relief; the severe horizontality of the incised line of the eyebrow parallels much in the tenth-century Khmer art, but the smooth transition from face to hair opens the possibility of a slightly later date, one in the eleventh century. Yet again, the absence of a context means that there is a limit to the conclusions that can be drawn. What it is important to realize is that in this period iconographic types were established across stylistic boundaries. The Buddha with left arm at the side was not a significant type in the later eleventh or the twelfth century, but towards the end of the thirteenth century it became popular again. The survival may be the result of cultural connections between Sukhothai and northeastern Thailand.
The naga-protected Buddha is the second iconographic type to consider. It had been established in Dvaravatl art in a period substantially earlier than the tenth century (pl. 25; pp. 95 and 97 above), and like the left-arm-at-the-side Buddha, it was borrowed into Khmer art, perhaps at about the same time—the mid-tenth century. The naga-protected Buddha quickly became, as the Beng Vien type did not, the chief form of the Buddha for the Cambodians. It appears at Phimai and is the central Buddha in the art of the Jayavarman VII period. The fact that the type became so firmly connected with Cambodian cultural hegemony may account for its eventual disappearance in the period after the thirteenth century.
This nearly life-sized sculpture stands isolated, with no other works identifiable as those by the same sculptor (or by his teacher, or his disciples). The belt, seen also in pls. 36A and 36B, is a tenth-eleventh-century feature; other ties to the same works include the central pleat of the undergarment, visible between the legs, the pinched torso (shared with the bronze, pl. 36B), and the excessive distance from the bottom of the lower edge of the robe to the bottom of the undergarment (shared with Wat Php Ta, pl. 36A). Of the three works, this large stone image has the least Khmer face; the soft modeling of the eyebrows is somewhat reminiscent of that in the late Dvaravati bronze, pl. 37, but here the head, raised on a high neck, bounded by firm jaw, has an individuality shared more by bronzes with a folkish character. It was not until somewhat later, at Phimai, that the Buddha with double gesture was renewed, transformed, and given new meanings with the addition of a crown.
Another iconographic type deserving mention is that of the standing Buddha with right palm raised, left palm lowered. The authoritative Buddhist images were increasingly, it would appear, ones that were Khmer in style.

Central Siam: a Geographical Perspective
It is not yet possible to tie this description of trends together with an archaeological record. The sort of archaeological evidence that is crucial in a period of cultural change—the sort that establishes when worship ceased at a particular temple or that a town was suddenly abandoned has simply not accumulated. Nor are there enough inscriptions. One potentially significant inscription, thought to have been found in Lopburi province, dates from 923 and lauds the then- reigning monarch Jayavarman. The Lopburi inscription (Th. 19) which conveys King Suryavarman I’s order of 1022 that the tapas of yogins and of Mahayana and Sthavira monks be offered to the king himself is the first solid indication of Khmer political control in the Chao Phraya basin (if indeed the inscription was not moved from elsewhere). Before then, there is plenty of evidence regarding cultural penetration but uncertainty about the political situation.
Rajendravarman (r. 944-68) left two inscriptions along the modern border, in the districts of Aranyaprathet (A. D. 941; K. 957) and Ta Phraya (K. 999). In both instances he seems to appear as a Saiva temple administrator (K. 958, st. 11), before his ascension to the throne. There may or may not be a connection between these activities, the western expansion of the kingdom and Rajendravarman’s victory over the Ramanya (above, p. 137). A lintel depicting Visnu at the center, probably from Rajendravarman’s time, comes from an unidentified temple in Wattana Nakhon district, just west of Aranya- prathet. Once the jump is made into the neighboring geographical region, that of the Prachinburi-Bang Pakong River basin, the situation is less clear—though potentially it is quite rich. Although the stone Buddha image in pl. 40 has here been attributed to Muang Si Mahosot, archaeological evidence from there and from Muang Khok Khwang, on the other hand, reveals little or nothing about the tenth or eleventh centuries.
The more northerly route of communication between east and west ran westward from the site of the early tenth-century Khmer- style Prasat Myang Kh*k (above, p. 120) to the Pa Sak River watershed and on to Lopburi. Perhaps in the tenth century there was more intercourse in this region than through Muang Si Mahosot. Near the banks of the Pa Sak River in Chai Badan district, there must have been activity; here was the Dvaravati site of Sap Champa, and at some point in time Khmer culture became established, as evidenced by a brick shrine (Prang Nang Phom H§m), an eleventh-century antefix (or possibly twelfth) from the site showing a standing female figure, and stray finds like a small eleventh-century naga-protected Buddha.
Lopburi had evidently been a Dvaravatl town of importance. The name Lava or Lavo appears on a Dvaravatl-period medal found near Nakhon Pathom, in a Khmer inscription of A. D. 611 (K. 557), and in one of the locally found Khmer inscriptions (Th.21), and it survives today as Lop-(=laba)buri. The innermost moat appears to date from Dvaravatl times. Excavations at Wat Nakhpn Kosa in 1986 brought to light stucco fragments—head of the Buddha, supporting dwarf, colonnette fragment, and molding with the bilobate motif—that may be plausibly attributed to the period around 700. The largest body of stuccos, discovered in 1937, on the other hand, appears to date from considerably later and may be no earlier than the oldest surviving “Khmer” monument—Prang Kh£k.
In style the sanctuaries are not far distant from Baksei Chamkrong at Angkor, built in the first half of the tenth century. The bricks vary in size—a characteristic of Dvaravatl practice, according to Anuwit Charernsupkul. Inside the sanctuaries bricks extend out from the walls, forming trefoil niches (pl. 39B). They suggest the presence of Pala influence—as can be seen by looking at pl. 38B—but painted niches somewhat similar in form have been found in a tenth-century Cambodian temple.
In and of themselves, Prang Kh£k and the Lopburi stuccos do not reveal anything about the local political situation; both an independent and a subservient ruler could equally well have made use of local craftsmen. If, on the other hand, an unfinished tenth- century lintel now in Ayutthaya did in fact come from Lopburi, perhaps activities there should be seen in an imperial context, the lintel being so high in quality. At any rate, inscriptional evidence of imperial control appears by the 1020s, during the reign of Suryavarman
I (1002/1011-1050)—if this inscription was indeed set up in Lopburi. An unpublished inscription from the king’s reign states that Lavapura (Lopburi) was at this time a jungle filled with tigers, more terrible in appearance than a cremation ground; an official, appointed “chief of the Mons,” was charged with restoring its former glory. The actual extent of Suryavarman’s military exploits in fact remains undetermined. It is possible that the laterite foundation known as San Sung dates from the Suryavarman period, or at least from the eleventh
century.
There are works of eleventh-century sculpture, two impressive naga-protected Buddhas among them, that are essentially cosmopolitan in style and are likely to have been imported, either in the eleventh century or later. The whole question of a local sculptural tradition in the tenth and eleventh centuries—indeed right up to the very end of the twelfth century—is an open one. One approach posits a continuous surviving series that begins with a “late Dvaravati” earth- touching Buddha, in which there are Khmer elements. According to an alternative view, Khorat-series sandstone images exhibiting an eleventh or early twelfth-century character are in fact no older than the late twelfth, though perhaps dependent on a body of older, wooden sculpture, now lost. This uncertainty gives double importance to two works, a Brahma (pl. 41B) and a pair of goddesses that probably date from the late eleventh or very early twelfth century. The medium is hard, dark stone; it is been proposed that both slabs were carved from what had once been a Dvaravati pillar. Here is arguably a local product. The facial type is somewhat like that seen in a small naga-protected Buddha from Suphanburi (pl. 41A), but there is a strong element of schematic simplification, especially evident in the planar treatment of the eye sockets.
Moving westward from Lopburi, the north-south riverine arteries are crossed: the Chao Phraya, the Suphanburi, and the Mae Klong. It may be too easy to conclude that at the old Dvaravatl centers there was disruption and displacement, that Mon inhabitants were forcibly resettled as prisoners of war or were decimated in a series of plagues—too easy because much of the evidence is merely negative evidence, and because certain kinds of artifacts like small Dvaravatl bronzes have not been adequately studied. The same pattern appears to hold for much more significant towns like U Thong and Khu Bua. At Nakhon Pathom, there was significant activity in the thirteenth century, and so the city was probably not abandoned, but works of the eleventh and earlier twelfth centuries have not been identified.
In some cases, it is possible to glimpse intrusive objects of Khmer workmanship. Among the bronzes found at Phong Tuk, for instance (though not excavated in situ), is a small tenth-century bronze image of the Buddha seated within an aureole. In other cases it is possible to point to sites that may have become settled in this period, like Ban Nong Chaeng (Don Chedi district, Suphanburi), at which a tiny stone eleventh-century naga-protected Buddha was found, on the order of the image from Wat Pfi Bua, Suphanburi (pl. 41A). Sometimes it is merely a matter of tantalizing gaps. The lobed frame on the votive tablet from Ban Samphao Lom (pl. 29B) on the
Suphanburi River in Doembang Nangbuat district, for instance, is relevant to designs seen on thirteenth-century bronze altarpieces. Two other somewhat related tablets have been found at the same site, as was a Pala bronze. There are hints, therefore, of continuities among Pala-influenced objects dating from late Dvaravati times onward. Much in the way of architectural remains has been destroyed, as at Ban Samphao Lom, however; and it will take both skillful archaeology and careful study to define the lineages that survived Khmer intervention.
Thirty kilometers north of Doembang Nangbuat district is Sankha- buri or Muang San, later an important center. An eleventh-century lintel here may have come from the site of Ban Dong Khon, which has both Dvaravati and Lopburi period vestiges. Sixty kilometers north lies a site in Tha Tako district with indications of eleventh- century activity, in the form of stone naga-protected Buddha images, plus Dvaravati ruins that would appear to somewhat pre-date the eleventh century. From there, another fifty kilometers to the northwest, along the course of the Ping River, brings one to Dong Mae Nang Muang, in Banphot Phisai district, Nakhon Sawan, where a significant inscription of A. D. 1167 was discovered (below, p. 163). The little-known monumental and sculptural remains suggest that Dong Mae Nang Muang was indeed a town at which Dvaravati traditions were maintained into a late period. There is evidence of contact with Haripunjaya. It is hard to know, however, just how late a stele with a standing Buddha, his feet clumsily oversized, may be. At any rate, if the Fine Arts Department reported no Khmer artifacts, the inscription itself is in the Khmer language, and so even here the pattern of Khmer penetration applies.

Phimai
Phimai, in Indic spelling bimaya or vimaya, was the name of the principal image at this great temple (pls. 42-45). Although no central Thai historical traditions include references to Phimai, it is, nevertheless, certain that Phimai occupied a seminal position in the religious heritage of central Thailand. It was at Phimai that the crowned Buddha became established as an iconic type, and the temple must be considered the prototype for the prang of central Thailand.
An inscription from the region dated 1066 proclaims a Vajrayana Buddhist doctrine, which it calls the Srlsamaja (perhaps in reference to the Guhyasamaja-tantra), involving worship of the five sugata (Jinas) plus Vajrasattva. This inscription, like the Phimai temple itself, combines Vajrayana elements with older local traditions. Dhanu, the author, records that he has installed nine Buddha images that his teacher, Dharanlndrapura by name, had restored. Dhanu does not doubt that he lives in Cambodia, and if the “protecting lord of the all-around lighted sphere” (Srisamantaprabhesvara) is really a reference to King Suryavarman I, it is he who is given credit for consolidating Buddhism in the kingdom.
Geographically, Phimai lies in a position peripheral to both Angkor, on one hand, and Lopburi and Ayutthaya, on the other. This eccentric position helps explain its dual role. Some of the stone carvers of Prasat Khna Sen Kev must have been subsequently taken to Phimai, where the rather more developed ornament is of equally high quality. Features of this ornament can be seen in the photograph of the distant cornice (pl. 43): the integrity given to each register; the deep undercutting; the precision and sharpness of the rhythms; the attention to the role of light and shade; and the opposition and balance of forms, as in the spreading and contracting pendent elements within the frieze, or the crosses and Xs within the quatrefoil band. Another development was the increasing prominence of figured lintels (pls. 44-45), and here, too, the designers were building upon earlier eleventh-century cosmopolitan tendencies. If the Mahidharapura dynasty did not come from Phimai, what accounts for the movement of the kingdom’s finest craftsmen to this eccentric location? The inscription on the main gate of the second enclosure provides a clue: in A. D. 1108 V. K. A. Sri Virendradhipativarman established an image of a Buddhist deity who is called the senapati, or general, of the principal image, the K. J. Vimaya (K. 397). The implication appears to be that Virendhradhipati served the reigning king, Dharanindravarman (r. 1107-1113), as general, just as the image he erected served the lord Vimaya. Virendradhipati was, therefore, the king’s loyal servant, but possibly at the same time an eminence grise more powerful than the king himself. He makes an appearance later among the generals in the procession of the southern gallery at Angkor Wat.
Presumably most but not necessarily all of the temple was completed when Virendra put up his inscription, probably in A. D. 1112. Although the figural style and the ornament at the temple should be viewed in the context of royal Khmer traditions, the design of the temple as a whole suggests the presence of other currents. Two significant features can be seen in pl. 42, which shows the temple after completion of the reconstruction in the 1960s, and both are important to subsequent Thai developments. One is the treatment of the re-entrant angles at the corner of the sanctuary: three corners (two in the photograph) are allowed to run from dado to cornice unimpeded. The second feature is the height of the moldings around the sanctuary: the dado carvings are almost as high as the first false window, and visitors at ground level are overwhelmed by the molding sequence. Neither of these features appears entirely explicable merely in terms of royal traditions. It may be possible that there was some sort of regional tradition, an independent development starting from Prang Kh£k in Lopburi. But surviving brick temples are not sufficient to demonstrate that this was the case, and perhaps outside influence, possibly from northern India, also played a role.
The religious imagery is complex, and it is not clear how the parts are all supposed to fit together. The pediment over the main entrance, on the south, features a dancing Siva (as at Phanom Rung), and the inclusion of the Tamil female saint Karaikkakkalamnaiyar indicates contact with southern India. On the other pediments and lintels of the exterior Saiva themes continue, as on the eastern pediment of the hall (pl. 42), where on the upper part of the outer pediment Siva can be seen with Uma on a bull, but scenes from the Ramayana are more prominent. Not all the reliefs have been identified. At the foot of a pilaster beside the main southern entrance a Tantric Buddhist figure appears, his feet on a corpse, his hands holding a vajra and a ghanta; he may be Vajrasattva in a secondary role as one of the sixteen vajra beings in certain esoteric mandalas. The fact that Phimai is a Tantric Buddhist temple is, therefore, not completely disguised by the imagery of the exterior, yet the Buddhist message lies in the interior. The south-facing main image, the lord Vimaya, is either lost or lying unidentified, but, like the deity of the innermost southern lintel, this image was probably a naga-protected Buddha. The other four interior lintels have Buddhist subject matter and may each involve conquering. The outer southern lintel depicts the defeat of Mara; on the western lintel (pl. 44) is a standing crowned Buddha at the center of a crowd of figures.
In surveying the Buddhist inscriptions of Cambodia, Jean Filliozat pointed out that many of the elements in Khmer Buddhism can be understood in a Saiva context. This is to be explained in part by the Saiva-like character of Buddhist Tantrism, in part by a Khmer tendency toward syncretism. How much of Phimai can be understood sheerly in Saiva terms is an open question. Dharanindravar- man I’s inscription at Phnom Sandak (A. D. 1110, K. 191) opens by stating that Siva, though unique, appears twofold, in knowledge (jnana) and activity (kriya). Perhaps at Phimai the images can be divided into spheres of uniqueness, knowledge, and activity. Near Prasat Khna Sen Kev (and Phnom Sandak), in the presumed home territories of the Mahidharapura dynasty is a little-known Buddhist temple that may have both stylistic and iconographic points of connection with Phimai.
One of the inscriptions most relevant to the temple is a fragmentary one found within the temple compound, bearing a date equivalent to A. D. 1041 (K. 953). Although it is unlikely that any of the main temple dates from that period, a brick and stucco ruin probably does. On one face of this inscription there is a stanza invoking Siva, who is both unique and multiple; on the other side is a verse praising the Buddha in comparable terms: though nondual, he has four bodies, as if in fear of the four Mara (catuskayas caturmmarabhayad iva). The inscription lends itself to the hypothesis that the main image, the lord Vimaya, is the nondual Buddha, and that his four bodies are represented on four of the lintels: on the outer southern lintel he conquers devaputra Mara, Mara as he appeared to the Buddha at the time of the enlightenment; and on the western, northern, and eastern lintels he conquers the remaining three Mara, evil (klesa), the aggregates that form the personality (skandha), and death (mrtyu). It is conceivable that the name Vimaya has a double meaning and is intended to convey both this uniqueness and this multiplicity: on the one hand it is “the one free from illusion,” but it may also be “the one manifesting various illusions.”
On the inner western lintel (pl. 44), the Buddha stands between two trees. He wears a metal belt with pendants, and older photographs show his head and diadem. His hands, long disappeared, presumably executed vitarka-mudra with both hands. To the Buddha’s right a man with Brahmanical hairbun crouches, his hand holding a bell (ghanta) in front of the tree. Many of the attendant figures hold vessels—some cylindrical, some in the form of half-seated animals, some in the form of a kundika. In a lower register there are three dancers and numerous musicians. The rhythmic precision and complexity in the dancing and moving figures, and the textural differentiation apparent in the relationship of the Buddha to the trees that frame him, are comparable to what can be seen in the ornament of the temple as a whole (pl. 43). The vessels held by the members of the court may allude to some sort of funerary rite. The Buddhist inscription of Bat Cum (A. D. 953, K. 268) says that the Buddha, who has exterminated Mara the enemy king by the detachment that rises from samadhi, just as fire rises from a funerary pyre, shines and rejoices as a supreme king (adhiraja) in the splendid palace of nirvana. The stanza pulls together the themes of the defeat of Mara and funerary rites; even more importantly, it presents an image—that of Buddha as supreme king in the palace of nirvana—that accounts for the royal attire.
An unfinished lintel at Phimai (pl. 45)—its original destination unknown—can be interpreted in such a way as to strengthen funerary associations of the western lintel and of the monument as a whole. The crowned Buddhas of the upper register should be identified as cosmic Buddhas, gathered temporarily at this spot from the different points of space. On the lower register is a group of human figures, some male, some female. The most important personage lies prostrate on the ground, his arms outstretched. A horizontal bar separates the two realms, the human and the transcendent. At the center of the lintel, however, there is a link between the realms, in the form of the naga coils that support a figure of the Buddha, an uncrowned Buddha. So the naga, it could be said, lifts the Buddha out of the worldly realm into the heavenly.
Beneath the naga there is a bird, which seems to be supporting the serpent, pushing its coils into the upper register. Then there is a vessel, perhaps of bronze, with an elaborate spout, and something coming out of the spout, something that might in fact be a flame emerging from a lamp, a flame that at the top is crooked, as it turns and touches the middle of the naga’s three coils. One possible interpretation is that the bird and the naga are imaginary but the vessel with its flame, like the important personage presenting it, is real. What is depicted is the lighting of a funeral pyre. The Buddha is really the soul of the deceased.
Such a highly speculative interpretation leads to the general notion that the naga can be considered a means of transport to a heavenly realm. The unfinished lintel—regardless of where it was originally intended to stand—would have complemented the funerary theme of the western lintel. This lintel, however, may embody other themes, important for the role of the royally attired Buddha in subsequent centuries. When George Credes discussed the lintel in 1923, he suggested that it be interpreted as representing a legend found in later Lao, Cambodian, Thai, and Burmese literature, that of King Jambu-patl, a heretic king who is finally converted by the appearance of the Buddha as king of a magically created city. In some ways the single stanza in the Bat Cum inscription seems to apply to the lintel more satisfactorily. At the same time, however, it makes sense to think of the Jambupatl tale—which is very much the tale of a con- quest—as also one which perhaps in an earlier form was part of the Phimai milieu. As the crowned Buddha spread as an iconic type in the twelfth century, it did not carry with it all the deities of Phimai’s esoteric Buddhism. But a story much like the Jambupatl tale must have traveled with the images of the Buddha in royal attire.
Even though the stanza from the Bat Cum inscription appears relevant to the Phimai lintel, to what degree Phimai Buddhism is descended from that of Bat Cum is not known. One issue is the matter of fourfold relationships, as found—apparently—among the lintels of Phimai, and triadic systems. The Bat Cum inscriptions (K. 266-68) invoke triads consisting of the Buddha, Vajrapani, and either Lokesvara or Prajnaparamita; to what extent the Buddhists who worshiped these triads should be distinguished from worshipers of the Beng Vien triad—Buddha, Prajnaparamita, and Lokesvara—has not be determined. Similarly, the connection between these systems, the later Phimai developments, and the sophisticated esoteric Buddhism found in the Cham inscription of An-thai (A. D. 902) is not clear.
A tenth- or early eleventh-century stele that may have been made in the Phimai region can be understood as mediating between threefold and fourfold systems. On one side is a naga-protected Buddha, on the other an eight-armed Lokesvara; on the narrower faces appear a four-armed Lokesvara, to the naga-protected Buddha’s right, and Vajrapani, to his left. It is possible, therefore, to take the four-armed Lokesvara, the naga-protected Buddha, and the Vajrapani as representing a standard triad. Beneath all these deities are female figures, and so the stele has a strong Tantric cast.
Why was the Dvaravati type of double gesture borrowed for the Buddha of the western lintel? If the associations with death and with quasi-resurrection as supreme king in the palace of nirvana are valid, then it may be that the image type was that of certain memorial sculptures important for familial reasons to the builders of the temple. (There are other lintels with standing crowned Buddhas, sometimes in a row of seven, but their original location is not known.) Archaeological excavation at Phimai indicates that in the upper layers Khmer glazed wares were mixed with the Phimai black ware that had been characteristic of the area since perhaps sometime in the early first millennium (above, p. 18). In other words, the building of Khmer-style temple at Phimai did not involve the displacement of local people or the eradication of the local culture. The Buddha of the western lintel is a tenuous but equally significant link with these regional traditions.
Bronzes recovered at or near Phimai include a Tantric Buddhist deity in the dance pose of pratyalidha, with vajrahumkara-mudra (pl. 46A). There are reasons for believing it to be a form of Trailokya- vijaya, the krodha mentioned in one of the Phimai inscriptions (K. 397). The fullness of the upper torso of this figure suggests that it is really a Baphuon-style figure, perhaps datable to about the third quarter of the eleventh century and therefore older than the major part of the temple itself. An interesting feature is the string necklace, which also appears on certain Buddhist dancing female partners that may well be the product of the same milieu. Perhaps the necklace type is derived from northern India and can be taken as evidence for the spread of Tantric concepts in the middle decades of the eleventh century, exactly at the same time Tantric masters were moving from the Pala monastic centers to Tibet.
Technical analyses of both the bronze alloys and of the clay cores may clarify the extent to which a local Phimai workshop had unique practices. Of especial interest is the question of mercury-gilded bronzes: where they were made, for how long a period, and whether the technique was introduced from Pala India, Java, or elsewhere. Some Buddhist bronzes which have been associated with Phimai appear to provide evidence for the persistence of a local tradition dating back some centuries. One of the bronze altarpieces that with some confidence can be placed very early in a sequence that was produced from the eleventh through the thirteenth centuries (and possibly later) is illustrated in pl. 46B. It was said to have been found at Phimai and may be assigned a date of ca. 1100. The general model is Pala: a stylized Bodhi tree appears at the top; the aureole is round-arched, bordered by a pair of moldings and flattened leaves; the standing Buddha is a Phimai transformation of the Dvaravati iconic type, with both of the Buddha’s hands in vitarka-mudra, but with a ribbed crown, necklace, and jeweled belt with pendants. The Buddha stands on a round pedestal, framed by a cut-out base, both resting on a lower supporting base having a middle frieze of supporting lions. A standing image with close stylistic ties to this Buddha, in the Walters Art Museum in Baltimore, was discovered to be made of an alloy of 80% copper, 22% tin, and 6% lead. The very high level of tin suggests that the workshop descended from that responsible for the Prakhon Chai bronzes and that the elevated amount of lead is be a feature distinguishing the local tradition from that of the capital.
The bronze workshops associable with Phimai may be said to have had a dual offspring, but in neither case can the details of the lineages be traced, and the story has to be picked up again at some point in the second half of the twelfth century. One tradition was that of Tantric images. A celebrated Hevajra, for instance, whose body is twisted in a manner much like that of Krsna on a Phimai lintel, came to the Bangkok National Museum from Battambang province in Cambodia. This may or may not mean that bronze- casters active at Phimai moved elsewhere after 1113. The other legacy consists of altarpieces, with either a seated or a standing crowned Buddha. The first tradition was Tantric, the second Hinayana. As the official art of the Suryavarman II period was predominantly Brahmanical, and the amount of Buddhist art produced is not known, the immediate fate of Phimai’s bronze workshops is obscure.

The Age of Suryavarman II
It was in many ways the political disunity of the kingdom that made Phimai possible. Jayavarman VI and Dharanindravarman may not have ruled over all of Cambodia, but their successor, Suryavarman II, who removed his great uncle Dharanindravarman from the throne in 1113, most certainly did. Once again Angkor became the focus of activity, and in creating Angkor Wat, Suryavarman outbuilt all his predecessors. This concentration of power was once again followed by a period of weakness at the center, however, for after Suryavarman II’s death sometime around 1150 Dharanindravarman II was unable to maintain the firm guiding hand of his predecessor. The relationship between center and periphery is exemplified by the missions from Lo-hu (Lavo, Lopburi) to China. The mission of 1115 must have occurred while Suryavarman was in the process of reuniting the kingdom, an attempt by Lavo to see if independence could be maintained through diplomacy. The mission of 1155 evidently took place after Suryavarman’s death; it may be understood as a declaration of independence.
In the case of Brahmanical shrines, the evidence suggests that 1113 was in no way a watershed. One of the inscriptions of Vat Phu, for instance, provides support for cultural continuity: it first describes activities of Jayavarman VI in A. D. 1103, and goes on to state that an image was erected in the vrah pran in A. D. 1132 (K. 366). Along the Angkor-Phimai route, the major construction was at Phanom Rung, a long-established sacred site. Whether construction of the main temple there spans the Jayavarman VI-Dharanindravarman- Suryavarman II periods, or dates entirely from the reign of Surya- varman II, however, has not been determined.
Prasat Khao Phanom Rung is a tower-sanctuary with antechamber, much like Phimai in plan but distinguished from Phimai by its Brahmanical orientation, its dramatic hilltop setting, and its elevated causeway and monumental stair. Many elements suggest stylistic continuity with the Buddhist temple; the entablatures at both sites, for instance, are similar in organization, and both are developments of the scheme at Prasat Khna Sen Kev.
Adding interest and complexity to the story of Phanom Rung is a group of inscriptions, some of which were uncovered in the course of the restorations that were carried out from 1977 until 1988. One inscription appears to record the deposit— possibly by Suryavarman II himself, including one of the dancing Siva, in a foundation stone. Another inscription was found in two sections, the top part (K. 384) observed in the nineteenth century, the bottom (K. 384 bis) uncovered in 1972. The author of this inscription was Hiranya, who in the Suryavarman II period erected a golden image of his father and teacher Narendraditya (K. 384). This Narendraditya, a cousin or nephew of Suryavarman II, initially had a military career and then retired to Phanom Rung as a Saiva guru adhering to Pasupata doctrines. Conceivably an inte­rior lintel at the temple depicts Narendraditya’s consecration as a Saiva hermit, flanked by the followers who have accompanied him from Angkor.
Prasat Phanom Rung may surely be considered Narendraditya’s temple, but whether it was established by him, for him, or in memory of him remains unclear, as is the question as to whether he was a man with deep roots in the area or something of an interloper at the time of his arrival. The sculptural imagery that supports the Saiva orientation is found on three key axial pediments: the first, at the eastern entrance gopura, depicts Siva Daksinamurti (as if Naren- draditya were an embodiment of Siva the teacher); the second, at the entrance to the antechamber porch, bears the dancing Siva; and the third, an upper pediment at the entrance to the antechamber proper, has been identified as an enthroned Siva and Uma blessing the asuras. Pediments on the southern and western faces also bear Saiva themes, and rows of rsis appear on the interior.
Below the dancing Siva of the principal pediment is a lintel depicting the Visnu Anantasayin (pl. 47), making a juxtaposition of the creation and destruction of the world. This lintel is at once powerfully tense, because of the poised serpent and the foliage, and gentle, because of the goddesses’ caressing hands. Visnuite subject matter, primarily in the form of scenes from the Ramayana, in fact dominates the remaining exterior pediments and lintels. Visnu Trivikrama also appears, as does Krsna—in combat, supporting Govardhana, and killing Kamsa. Indra is shown on a few lintels; the Mahabharata makes a very limited appearance; and the directional deities can be seen on antefixes. There are echoes of Phimai, in the presence of the dancing Siva on the principal axial pediment and in the role of the Ramayana. Open questions include whether the inner lintels were intended to have a privileged function (as at Phimai), how to understand the relationship to contemporary imagery at Angkor, and whether the combination of Saiva subjects with the Ramayana scenes is intended to allude to the consecutive stages of Narendra- ditya’s career.
The impressive naga bridge or elevated terrace at the foot of the grand staircase leading to the Phanom Rung sanctuary may be later than the sanctuary itself. The terminating nagas have crowns (or borders) consisting of narrow registers of shallowly incised decor. This feature also characterizes nagas found at the Angkor Wat-style temple of Beng Mealea. As can be seen in pl. 48A, the terminating nagas of the pediments on the sanctuary tower also have these bands of shallow decor (as do the naga antefixes), but they are sandwiched between the naga heads and the robust border of leaves. At Phimai only a molding separates the leaves from the heads. Such details hint at a chronological sequence, and at a stretched-out but unbroken period of construction.
Two other important monuments apparently also belong to the decades around the beginning of Suryavarman’s reign. They share a significant motif—the entrance-flanking pillars displaying on one face a guardian, on the other a devata garbed in a manner associated with the Angkor Wat style. At the three brick towers of Prasat Prang Ku (Sisaket province) lintels evoke (or did evoke, before most of the stone elements were removed) the earlier eleventh-century style but include dancers and architectural elements of a sort found at Phimai. As for freestanding Brahmanical sculpture in northeastern Thailand during the Suryavarman II period, very likely certain local workshops, with regional characteristics, remained active.

Buddhist Traditions
If all the resources were flowing toward Angkor, where there was a demand for craftsmen, and if the imperial art of the time was overwhelmingly Brahmanical, then there may in fact be little provincial Buddhist art to describe during the reign of Suryavarman himself. That may be an accurate picture, but it is by no means a certain one, for it depends on judgments of chronology that are simply too subtle.
A lintel depicts the assault of Mara, but the army is dispersed among scrolling foliage, in a manner typical of other lintels of the Angkor Wat period. The members of the army, however, are in style similar to figures on the maravijaya lintel of Phimai, or to the figures on a bronze image base that shows the army attacking on one side of the Buddha and paying homage on the other. If that is the case, various schools of Buddhism must have been developing within Cambodia as a result of direct contacts abroad.
Phimai did not give rise to other Tantric Buddhist temples, at least not in the Suryavarman period. Some significant Buddhist stone sculptures have been found in the Northeast, but their chronological position and relationship to architectural remains are hard to fix: one is a Lokesvara found at Prasat Ta Myan Thom; another is a crowned naga-protected Buddha from Ku Santarat, Maha Sara- kham. Such artifacts tend to have a floating character that is even more of a concern in looking at central Siam—not merely in the obvious case of bronzes, but with stone sculpture as well. In the stele in Suphanburi from Wat Khao, said to be made of green stone (pl. 52), there is a sophisticated unfolding of forms—the forehead wrapped by the diadem, the usntsa that spreads out to fill the diadem, the hood that frames the head and crown while also filling the space of the niche, the undulant curves of the “architectural” parts matching in rhythm the curves of the “natural” serpent hood. If this is a work that was carved locally, it suggests the movement of a highly trained sculptor to an unidentified urban center somewhere in the Suphanburi region around the early twelfth century. It rests isolated as a work of stone sculpture, and although it is possible to imagine it as standing at the head of a series of twelfth-century bronze images produced in the Suphanburi region, no sequence of bronzes that would span the entire century has ever been assembled.
The second sculpture is the exquisite sandstone naga-protected Buddha from Ta Phraya district, Prachinburi province, along the modern border (pl. 53). This is a region so long integrated into the Khmer kingdom that there is no point in looking for regional features. Other differences may be considered personal ones: the sculptor of the freestanding image is less concerned with making the parts fit together in an intimate way, and so the usntsa is narrower, the shoulders squarer, and the eyebrows have a degree of relief. The result is an image of considerable individuality.
The third naga-protected Buddha, seen in pl. 54, is much better known. It makes the first two images look quite similar. The coils are much thickened; the scales and the necks now vie with one another on an equal basis; and the medallions on the neck are halved—a “Lopburi school” feature of significance. A lengthening process seems to have occurred, affecting not only the usntsa and the earrings but the face as well, bringing a loss of the sense of spherical volumes, and an increase in the degree to which facial features can be perceived as separate units. Unfortunately, it is not known where this image, which was brought to Bangkok from Ayutthaya, was made. Regardless of how it stands to the first two naga-protected Buddhas geographically, chronologically it must stand at a certain distance, and it is more likely to date from the period following the reign of Suryavarman II than from the reign itself.
A certain number of sandstone images of the naga-protected Buddha bearing crowns and more-or-less conforming to the Angkor Wat-style type have been found in Lopburi. They too might date from the time following the death of Suyavarman II, especially if this was the period in which the Mahathat was conceived and started to take form (p. 223). It has also been suggested that stone workshops were not active in Lopburi until the 1190s, when stone images were carved based on lost wooden examples of the previous decades. At any rate, the production of Buddhist bronzes following Angkor Wat traditions in the Suphanburi and Lopburi regions during the second half of the twelfth century must be presupposed. One example might be the bronze naga-protected Buddha seen in pl. 55, with upper- torso modeling reminiscent of the early Angkor Wat style (pl. 53) and facial proportions something like those in pl. 54, but with three tiered coils that gradually swell more dramatically than those in pl. 54. A mat has been placed upon the topmost coil, and the crowning pedestal is ornamented with medallions.
Phimai can be connected architecturally to the central plains via Si Thep. As Si Thep was—apparently—a constantly occupied town, it has the potential of yielding a continuous record from Dvaravati times into the thirteenth century. Be that as it may, one of its two best-preserved sanctuary-towers, known as Prang Si Thep (pl. 50A), must be brought into a relationship with Phimai, for it develops certain elements that had come into prominence there: one is the configuration of the shaft of the body of the temple, with multiple re-entry angles rising to the main cornice; the other is the role of basement stories, which lift the sanctuary high off the ground. (Of course Prang Si Thep is considerably smaller than Phimai.) Prang Si Thep is built of both laterite and brick, the latter, according to Anuwit Charern- supkul, recalling in technique that of older structures such as That Phanom and Prang Kh£k, Lopburi. It is a monument that can be seen as a precursor of Wat Mahathat in Lopburi (pl. 51), which is similarly characterized by a shaft of multiple re-entrant angles and a basement of even greater height. The connection gives rise to the speculation that the Lopburi Mahathat originally took shape in the twelfth century, perhaps in the decades following Suryavarman’s reign. The stucco lintel incorporated into the fabric of the Mahathat (pl. 50B) can be understood as dependent on Si Thep stone lintels of the period (either the late eleventh or the first half of the twelfth century), sharing with them the unusual trait of the presence of three tiers of scrolling leaves below the horizontal branch.
The period following the death of Suryavarman II around 1150 was crucial for subsequent developments in Thailand, for it was yet another period of weakness at Angkor. If the Lopburi Mahathat, with its stucco lintel, took form in this period, then it is easy to understand why the monument should have become a pivotal point of focus in the thirteenth century. Even when the post-Suryavarman period is filled with specific concrete evidence—a key inscription— rather than with mere speculation, however, the kind of history that emerges is shadowy and murky. On Sunday, February 5, 1167, the great king {maharaja) Asoka presented gifts to a relic (brah sariradhatu, for sariradhatu) installed in a monument at Dhanyapura, a town now known as the site of Dong Mae Nang Muang in Nakhon Sawan province, where the inscription describing the gifts was discovered (K. 966). This relic had the same name—Srl Dharmasoka—as the king himself. What did this mean? Perhaps that the king gave his personal name to a relic of the Buddha, or perhaps that the relic consisted of the remains of a former king, also called Asoka, which his successor installed at Dhanyapura. There are several good reasons for favoring the latter interpretation. Dvaravatl ceramic reliquaries or urns found in Nakhon Sawan province (fig. 17c) and the presence of skeletons by stupas at Dong Mae Nang Muang both suggest that stupas served a funerary function, and chronicles from the peninsula provide evidence for the use of Dharmasoka as an inherited title.
One face of the inscription recording the gift is in Pali, the other in Khmer. On the one hand, the use of Khmer provides evidence for Cambodian cultural penetration; on the other hand, the content of the inscription and its archaeological setting argue for the continued strength of indigenous political and religious institutions. Except for the language of the inscription, Dhanyapura was apparently largely untouched by Khmer culture.
Haripunjaya is another possibility, as is Nakhon Si Thammarat, where the name or title Dharmasokaraja was in use. It may also be that a Dvaravati-successor kingdom was able to establish itself in this period, with a capital at Nakhon Pathom or some­where in Suphanburi province.
At one point in the Nakhon Si Thammarat chronicles, the local Asoka is able to recover relics long buried in the earth. His legacy from the Buddha, in other words, consists not just of the relic but of the right to uncover the relic. Its possession by the king gives him a right over the soil, a right that has been willed to him by the Buddha. The king’s power—over the soil, over the local spirits, and over the populace—derives from circumstances that situate him clearly both in time—at a certain distance from the historical Buddha.

Fragment of a lintel, from Chanthaburi province. National Museum, Bangkok

Fragment of a lintel, from Chanthaburi province. National Museum, Bangkok