Monumentality in the Mekong Delta: Luminescence Dating and Implications

Brick monuments are one hallmark of emergent political complexity across much of mainland Southeast Asia during the first millennium AD. They first appeared in association with Indic statuary, large nucleated settlements, and the region's earliest states. Chronometric research in the Mekong delta on the earthen and brick wall that surrounds Angkor Borei (southern Cambodia) provides some of the earliest secure dates for brick architecture in the region. This paper presents preliminary results of the dating program at Angkor Borei and examines the utility of luminescence techniques for dating historic period monuments across mainland Southeast Asia. Being Earth and the Word, the brick is a goddess: ‘To thee, O Goddess, Brick, let us sacrifice with oblation’ (Kramrisch 1976:105) Brick construction technology appeared throughout much of mainland Southeast Asia by the mid-first millennium AD, and is associated with the earliest states in the region. Populations in these early states used bricks to construct encircling settlement walls, to erect religious structures for housing Indic deities (Parmentier 1932:187), to create Buddhist caitya or stupa shrines within and around their settlements, and to construct mortuary monuments that contained cremations. Given the large number of brick monuments in India and Southeast Asia, the cultural significance of the monuments themselves and their spread through the region and the difficulties of obtaining absolute ages for the bricks and brick structures using other means, a huge need exists to date bricks directly that were used in these public works. This article focuses on dating bricks used to construct first millennium AD monuments, shrines, and other public works. Several persistent questions concerning the earliest brick technology remain unanswered, including the range of construction uses, the linkage between this technology and other concomitant organizational shifts, and the origins of this particular technology. None of these questions can be answered, however, without a solid grasp of the time-space systematics. The Mekong delta (southern Cambodia and southern Vietnam) is an ideal setting for studying Southeast Asian’s early brick technological tradition. Not only was the delta a crucible for social and political complexity during the early to midfirst millennium AD (Hall 1982, 1992; Stark 1998, 2003; Stark and Bong 2001): the level of archaeological activity on this period has accelerated in the last three decades. Until recently, scholars dated brick monuments throughout the lower Mekong basin through comparison of their art and architectural styles with those found in South Asia (e.g., Parmentier 1927, 1932), or through dating inscriptions on stelae that were incorporated into the structures. Each of these approaches has generated a coarse-grained chronology that requires refinement. More recent efforts to use chronometric dating (e.g., Dao 1993) still concentrate on either associated statues of construction fill, rather than on the monument itself. Obtaining direct dates for a Southeast Asian brick technological tradition sheds lights on regional developments, on processes of technology transfer, and on historical trajectories in Southeast Asian and South Asian brick technological traditions. We concentrate first on the earliest appearance of brick as a construction medium in South and Southeast Asia, whose use in secular construction dates to the late first millennium BC in South Asia. We then turn to evidence for the earliest use of brick in religious structures across the region, and discuss data from research through the Lower Mekong Archaeological Project. The current evidence suggests that secular brick construction preceded religious brick construction, and that this former technological development parallels, rather than lags behind, patterns described for South Asia. PATTERNS AND POSSIBLE ORIGIN AREAS Empirical Patterns across the Region Residents in each of mainland Southeast Asia’s first millennium states used brick to construct both secular and ritual structures. Chinese accounts of “Linyi,” for example, describe the introduction of Chinese technological traditions like defensive settlement walls and ditches by the Chinese merchant adventurer Fan Wen in the AD 320s (Southworth 2004:220). The earliest ritual brick INDO-PACIFIC PREHISTORY ASSOCIATION BULLETIN 26, 2006 111 traditions in the region date between the 5th and 8th centuries AD, from the Cham of central Vietnam to the Pyu in Burma’s Dry Zone. In central Vietnam, the earliest preserved Cham brick and stone plinths (bases and platforms) at Tra Kieu and Po Nagar probably date from the 4th-5th centuries AD (W. Southworth, pers. comm., 2005). Brick construction techniques appear later in the archaeological record of central and Northeast Thailand. Some scholars argue for the existence of Buddhist monuments as early as the third or fourth centuries AD in the Dvaravati region (e.g., Indrawooth 2004:138). Convincing evidence for the use of brick and laterite in stupa and caitya structures in central and Northeast Thailand, however, is dated to the 7th century AD, based on art styles and paleographic analysis (Higham 2002:256-257, 264265). These multiple lines of evidence support a 7th-8th century starting date for brick construction techniques in central and Northeast Thailand. Brick architectural construction is also associated with the Pyu of Myanmar’s Dry Zone, and also in southern Myanmar in the early centuries AD (Moore 2004:3). The conventional Pyu date range runs from the 1st 9th centuries AD, and scholars now agree that settlement occupation began before the construction of the encircling 2-5 meter thick brick walls that are a Pyu hallmark. Fingermarked bricks are ubiquitous on Pyu sites, and were used to construct settlement walls, in monastic buildings, and public buildings. Pyu bricks have never been dated directly, but radiocarbon dates associated with architectural beams from a brick-walled public hall at Beikthano range from the 1st – 4th centuries AD (Aung Thaw 1968:20, 23, 62). At the Pyu site of Halin, radiocarbon dates from structural timber and from two gateways range from the 1st – 8th centuries AD (Myint Aung 1970:57). Most Cambodia scholars agree that brick monuments appeared by the beginning of the pre-Angkorian period (c. AD 500-802), and functioned primarily as repositories for Indic deities (Parmentier 1932:187). Funanese religious architecture may have consisted of brick foundations with wooden superstructures (Parmentier 1931:143), and their construction likely began after the mid-sixth century AD based on their absence from Chinese dynastic annals (Malleret 1962:349). Brick remained the primary construction medium until the late 9th century AD when both sandstone and laterite were added to the construction material repertoire. The use of brick continued throughout the Angkorian period and did not disappear completely until the end of the 11th century, and re-emerged as a building medium in the late 13th and early 14th centuries (Boisselier 1966:45-46). The highest density of preAngkorian inscriptions derives from Southern Cambodia’s Mekong delta, and the delta’s early brick tradition has received the most archaeological research attention in mainland Southeast Asia. French scholars like Jean Boisselier (1966) assigned this architecture to the end of the 7th century AD, based on related epigraphic and statuary evidence. In the Mekong delta, civil war led to a three-decade research hiatus after Louis Malleret’s (1959, 1960, 1962) pioneering work at the Vietnamese site of Oc Eo. Vietnamese archaeologists resumed work on “Oc Eo Culture” sites in 1979 (Vo 2003). The Lower Mekong Archaeological Project (LOMAP) has worked in southern Cambodia since 1996 (Stark et al. 1999). A FrenchVietnamese cooperation program (involving the Institute of Social Sciences in Hô Chi Minh City and the École Française d'Extrême-Orient) worked intensively at the Oc Eo sites in southern Vietnam between 1997 and 2002. Their work concentrated on five areas in the vicinity of the modern pagoda of Linh Son (lower eastern slopes of Ba The), including brick architectural features (Manguin and Vo 2000). Architectural and statuary evidence suggest that Vaishnavite and Buddhist cults became prominent in Mekong Delta sites during the 5th century. Often built upon foundations of large river boulders, few of the area’s earliest brick structures remain in even fragmentary form. Three excavated architectural sites have been dated from southern Vietnam, and Luong describes them as “urban agglomerations” (Luong 2000:63). They are Ba The/Oc Eo (c. 5th – 6th centuries AD), Nen Chua (3rd – 6th centuries AD), and Go Thap (4th – 5th centuries AD). Almost two-thirds of the more than 50 stone, wood, and bronze statues that have been recovered from Vietnam’s Mekong delta represent Buddha (Luong 2000:66). A total of four wooden Buddha statues have also been radiocarbon-dated (dates range from c.AD 300 – 600 [Dao 1993:116]). Numerous dates have been obtained from fill associated with tombs, brick monuments, and “Oc Eo Culture” sites (Dao 1993). In sum, research in southern Vietnam emphasizes a 5th – 7th century AD florescence in brick architectural construction (Manguin and Vo 2000; Vo 1998, 2003). Possible Origin Areas Identifying potential geographic source areas for the Southeast Asian brick tradition would greatly enhance our ability to evaluate competing interpretations of the timing and relative influence of various South Asian regions on what is now mainland Southeast Asia. Early to mid-first millennium AD sites in most of Southeast Asia’s early “core areas,” however, lack sufficiently intact brick architecture to warrant systematic comparison with specific South Asian culture areas. This is especially true in the Mekong delta, most of whose brick monumental localities were reduced to surface or subterranean foundations by the time that Lunet de la Jonquiere published the results of his architectural survey (1901, 1902-1911). Today, partially intact buildings are exceedingly rare in either Vietnam or Cambodia (also see Trinh 1996: 122). Given the paucity of architectural remains, most scholars have relied instead on comparisons of art traditions associated with temple lintels, or with statuary that the structures were built to house. Chinese accounts, as well as art historical evidence, suggests the co-existence of Hinduism and Buddhism across early Southeast Asia STARK, SANDERSON AND BINGHAM: MONUMENTALITY IN THE MEKONG DELTA 112 (e.g., Bhattacharya 1997:36). While Buddhism seemed to dominate the Pyu (Myanmar) and the Dvaravati (central Thailand) cultures, the Hindu religion generated a deeper material imprint in southern Cambodia and Vietnam than did Buddhism. Elizabeth Moore notes that brick finger-marking, a Pyu hallmark, has parallels in Bihar, Uttar Pradesh and Kapilavastu, and was associated with the adoption of Theravada Buddhist practice in Myanmar and Thailand (Moore 2004:3). To the east into the Dvaravati homeland (and slightly later in time), Indrawooth (2004:139) argues that Dvaravati art bears affinities with (1) Nagarjunakonda in the Krishna-Godavari region of (Andhra Pradesh) when it was under Ikshvaku rule between AD 200-350 (p. 138) and (2) Mahayana Buddhism in Deccan India, and particularly by the Gupta and Vakataka rulers (c.AD 300-600). In what is now Cambodia, pre-Angkorian art styles bear an even more mixed South Asian/Sri Lankan imprint. The Phnom Da art style reflects Mathura and Gupta artistic traditions (Boisselier 1989:31; Dalsheimer 2001:44), and shares similarities with statuary found throughout Peninsular Thailand and Malaysia. Numerous wooden, stone and bronze Buddha statues recovered from the Mekong delta resemble 6th century Amaravati traditions from Tamil Nadu (Boisselier 1966:266-267; Luong 1995). Boisselier (1966:60) concluded that the substantial diversity documented in pre-Angkorian architectural traditions reflected diverse sources of inspiration. The unresolved debate over notions of “Indianization” (see Mabbett 1997 and Bellina and Glover 2004 for recent discussions) often obscures the importance of examining extant archaeological information on early brick technological traditions from both South and Southeast Asia. Comparing the earliest appearance of post-Harappan brick settlement boundary walls and brick religious architecture in these two regions provides one empirical tool for evaluating the relative timing of such processes and the historical implications of this information. The earliest stratigraphically dated brick temple in an archaeological context dates to the first century AD (Haertel 1993:86; Ray 2004a:348-349), although earlier temples with brick plinths and foundations may have been constructed a few centuries earlier. The earliest surviving brick temples date from the 4th 6th centuries AD during the Gupta period (Kramrisch 1976:101-102; Ray 2004b: 351), while stone architectural traditions began to flourish after the 4th century AD (Michell 1988:18; Ray 2004b:350). These early temples shared sacred space with a variety of domestic, local and regional cults, and coincide with the earliest well-dated Hindu images and iconography (e.g., Michell 2000:42-43). The manufacture and usage of brick for Hindu temples required particular protocol and bore specific religious connotations: bricks form the “body of the sacrifice (yajñatanu)” (Kramrisch 1976:102). Areas of southern India that flank coastal Bengal like Tamil Nadu and Andhra Pradesh were particularly important source regions for the spread of Buddhism through the Bay of Bengal and into mainland Southeast Asia (Bellina and Glover 2004; Ray 2005). Secular brick coastal structures have been reported from the 3rd century BC, based on brick dimensions (Gaur et al. 2005). Brick Brahmanical shrines appeared as early as the first century AD across much of southern India (e.g., Sarma 1983; Krishna Murthy 1983: 65). In this region, Brahmanical temples did not appear before the 4th 5th centuries AD and the earliest well-dated religious structures date to the 7th century AD (Ray 1997:45-46). These comparative data form an essential foundation for studying the timing of the earliest secular and ritual brick construction in the Mekong delta. Comparative art styles suggest a 3rd 7th century AD date range for temple construction, while South Asian evidence for brick religious structures begins in earnest by the 2nd century AD. However, the lack of directly-dated brick from any site across mainland Southeast Asia is a major impediment to understanding this process. Chronometric research on brick construction features in and around Angkor Borei (southern Cambodia) provides some of the earliest secure dates for brick architecture in the region. This paper presents preliminary results of the dating program at Angkor Borei and its surrounding monuments, and also examines the utility of luminescence techniques for dating historic period monuments across mainland Southeast Asia. Figure 1. Locational Map of Angkor Borei in southern Cambodia. Reprinted with permission from the University of Hawaii Press from Figure 1, p. 52 in "A New Date for the Phnom Da Images and Its Implications for Early Cambodia." by N.H. Dowling, Asian Perspectives 38(1):51-61. INDO-PACIFIC PREHISTORY ASSOCIATION BULLETIN 26, 2006 113 RESEARCH AT ANGKOR BOREI Work by LOMAP, in collaboration with Cambodia’s Ministry of Culture and Fine Arts, concentrates in and around the southern Cambodian site of Angkor Borei, in Takeo Province (Figure 1). The archaeological site of Angkor Borei has been the focus of art historical and historical interest since the late 19th century (Aymonier 1901), because of its monumental wall, its collapsed brick structures, its artifacts, and its associated Phnom Da art style, the latter of which represents the earliest Khmer art tradition. Many historians, following Chinese accounts (Pelliot 1903), believe that Angkor Borei was an inland capital of Cambodia’s earliest civilization, which Chinese visitors from the 3rd 6th centuries AD called Funan (Coedès 1968; Wheatley 1983). Previous publications have reported on LOMAP research findings at the site of Angkor Borei; these focus on the site’s size, configuration, and occupational history (Stark et al. 1999; Stark 1998, 2003). Also notable are the existence of brick monumental features (both structures and a surrounding wall), and a range of satellite settlements that were linked to Angkor Borei by artificial and natural waterways. Dating brick monument samples from Angkor Borei is important for understanding the settlement’s developmental history, its relative influence at different points in time, and the origins and age of stone sculptures that were placed within these monuments and have thus far been the exclusive purview of art historians. LOMAP LUMINESCENCE DATING PROGRAM AT ANGKOR BOREI Twelve brick samples from six localities within the Angkor Borei site have been sampled as part of the LOMAP luminescence dating program thus far. Brick samples from two types of primary contexts were sampled (i.e., secular [perimeter wall] and sacred [shrines and temples]), and three phases of the brick luminescence program have thus been completed: (1) a 1997 pilot luminescence study of two bricks from collapsed monuments at Angkor Borei, including one that LOMAP members investigated in 1996; (2) the 1999-2003 focus on the perimeter wall surrounding Angkor Borei; and (3) the 2004 work on collapsed brick monuments within Angkor Borei and its general catchment area (Sanderson and Bingham 2004a, 2004b). Analytical work has involved both optically stimulated luminescence and thermoluminescence. These samples were analyzed using a Single Aliquot Regenerative (SAR) Optically-Stimulated Luminescence (OSL) procedure on extracted quartz grains. In-depth contextual, methodological, and analytical descriptions have been presented in previous reports (Feathers 1997; Sanderson and Bingham 2004a, 2004b), and are not discussed in this article. We begin with the Angkor Borei perimeter wall, which Groslier (1935:491) first described, and which has been the focus of continuing LOMAP field investigation since 1996. Previous trenching activities indicate that at least some sections of the wall were 2.4 meters wide and 4.5 meters tall, that the wall’s dimensions may have varied from one point to another, and that the wall was constructed using core-veneer methods (Stark et al. 1999:2223). Angkor Borei wall bricks conform closely to the preAngkorian brick technology described by Jean Boisselier (1966): they incorporate rice chaff rather than mineral temper, and were made in relatively standardized sizes (c. 10 x 20 x 40 cm) that parallel the pre-Angkorian bricks that Boisselier (1966:46) recorded (i.e., from 25 x 14 x 5.7 cm to 31 x 16 x 8.5 cm). Inclusion of organic, rather than mineral, temper may be indicative of local manufacture, while the standardized dimensions could indicate the use of mold construction techniques. Table 1 presents findings of LOMAP brick luminescence dating thus far, and we concentrate first on Angkor Borei’s perimeter wall. In sum, we sampled four bricks collected from each of the two wall sections and found that individual OSL ages from the bricks ranged from 170±150 BC to AD 420±120, with individual errors of the order of 5-10% of age. Results from AB 8, an artificial wall cut made by a farmer to drain his fields, provided a mean age of AD 65±90. Samples recovered from AB9, the southeastern most wall remnant (but not the wall’s original southeastern corner), provided a mean age of AD 270±90. Subsequent radiocarbon dating of a carbonized rice grain from this latter location yielded a calibrated date of AD 240 420, which closely parallels dates produced from the seven luminescence samples. These conflicting results suggest either that we are underestimating luminescence errors for some reason, or that the brick ages are not synchronous. The local averages of AB8 and AB9 sections are themselves internally consistent, however, and the AB9 radiocarbon date accords well with the luminescence dates. Interestingly, the mean age of AB8 and AB9 are more probably different than the same; we interpret these results as indicating that the Angkor Borei wall contains bricks of more than one age. Although LOMAP crew members have documented no fewer than 15 collapsed brick monuments within the walled site of Angkor Borei, few are sufficiently intact to merit documentation and sampling for dating. One of these, called “AB-5” (Stark et al. 1999:22-25) was sampled in 1996 and luminescence dates proved problematic (Table 1). Feathers (1997) suggested that brick samples from this temple may not have been fired, and recommended reliance on optically-stimulated luminescence techniques. Although bricks sampled from multiple locations in 2001 and 2004 by the Scottish Universities Environmental Research Centre (SUERC) laboratory suggest some firing activity, Feathers’ recommendation of optical techniques proved useful. A second brick building from Angkor Borei, now dubbed Ba Srê (previously “Vishnu temple” [Stark et al. 1999:26-27]), was also sampled in 1996, and has produced more reliable dates than “AB-5” (Table 1). Local authorities sanctioned the dismantling of this feature (which reportedly included some bulldozing) following STARK, SANDERSON AND BINGHAM: MONUMENTALITY IN THE MEKONG DELTA 114 Table 1. Sampling details for all bricks collected during LOMAP 1997-2004. Locality Context description (& site) Brick label Lab no. Sample description Luminescence Lab A SUTL1560 Central brick sample. SUERC (Sanderson and Bingham) B SUTL1561 Brick sample adjacent to A. SUERC (Sanderson and Bingham) Phnom Toch Very disturbed structure in eastern sector (Angkor Borei). C SUTL1563 Brick sample from one course below samples A and B. SUERC (Sanderson and Bingham) A SUTL1565 Central brick sample and associated soil. SUERC (Sanderson and Bingham) B SUTL1566 Central brick sample. SUERC (Sanderson and Bingham) Ta Tos Moated mound in southern sector (Angkor Borei) C SUTL1567 Peripheral brick sample and associated soil. SUERC (Sanderson and Bingham)