Co loa an investigation of vietnam’s ancient capital (2010

Nam C Kim1 , Lai Van Toi2 & Trinh Hoang Hiep2 History, legend and memory have long pointed to Co Loa, an earthwork enclosure outside Hanoi, as the seat of an indigenous power that gave identity to the people of the Bac Bo region, north Vietnam Survey, excavation and a set of radiocarbon dates now put this site on the historical map The main rampart of the middle circuit was built in the later centuries BC, before the coming of Han Imperial China Nor was this rampart the first defence The authors show the potential of archaeology for revealing the creation and development of a polity among the prosperous people of the Dongson culture Keywords: Vietnam, Co Loa, Iron Age, Han dynasty, rice, Dongson drum Introduction Mainland Southeast Asia is an excellent and under-utilised region for comparative studies of early state formation, with great potential for examining changing spatial configurations of human-environment relations through time (Stark 2006: 422) Ongoing archaeological and historical research on complexity can particularly benefit from a growing body of Vietnamese data from periods of early ranked society, information that can contribute to the mounting evidence for how and why small-scale, egalitarian communities developed into complex societies in pre-modern Southeast Asia In this paper we present the results of investigations at the abandoned fortified settlement of Co Loa, which potentially forms the node of such a polity in the Red River delta near the Gulf of Bac Bo, north of Hanoi We offer a dated sequence of construction and advance the case that this sequence signals the rise of an indigenous state society in north Vietnam, one which only later came under the influence of Han China Department of Anthropology, University of Wisconsin-Madison, 5240 W.H Sewell Social Science Building, Madison, WI 53706, USA (Email: nckim2@wisc.edu) Vietnam Institute of Archaeology, 61 Phan Chu Trinh, Hanoi, Vietnam Received: 14 September 2009; Accepted: 23 November 2009; Revised: 19 January 2010 ANTIQUITY 84 (2010): 1011–1027 http://antiquity.ac.uk/ant/084/ant0841011.htm 1011 Research Co Loa: an investigation of Vietnam’s ancient capital Co Loa: an investigation of Vietnam’s ancient capital Historical context The conventional reconstruction of the protohistory of the northern Vietnam region during the second half of the first millennium BC benefits from various sources, including Vietnamese oral traditions and Chinese historical records (O’Harrow 1979; Taylor 1983; Tessitore 1989) These two sources, however, offer contradictory accounts Chinese historical texts suggest that Vietnamese civilisation was a by-product of Chinese colonisation, describing how the imperial Han began colonising the region in 111 BC, consolidating control during the first century AD Not surprisingly, Han texts maintain that agricultural, metallurgical and political sophistication emerged among the local barbarians in Bac Bo because of imperial annexation, generally denying in situ and indigenous cultural development (O’Harrow 1979: 143–4) In contrast, oral traditions hold that Bac Bo was the nucleus of an indigenously developed Vietnamese civilisation with powerful kingdoms ruling over vast populations before the arrival of the Chinese (Taylor 1983: 3–23; Tessitore 1989: 36) In particular, Vietnamese chronicles describe an indigenous Au Lac polity centred at Co Loa during the third century BC In approximately 258 BC, a man named An Duong Vuong (also known as Thuc Phan) purportedly overthrew the Van Lang polity, consolidated power over the local communities with the establishment of the Au Lac Kingdom and constructed the fortified citadel known as Co Loa as his capital (Taylor 1983: 19–21) This event significantly predates the documented solidification of Han control over the region, which occurred when Han general Ma Yuan quelled a Vietnamese rebellion in AD 43 and local Dongson warrioraristocrats were incorporated as a Han Empire province (Higham 1989: 202, 290–91) At that point, a loosely imposed Han tribute system was superseded by a full Han administration (Higham 1989: 291) Compounding the uncertainties are textual accounts describing the overthrow of the Au Lac polity in approximately 170 BC by Zhao Tuo, a former Qin general, who then incorporated Co Loa’s area into the rule of the Nam Viet (or Nan Yue) polity (Pham 2004: 202) Given these conflicting historical reconstructions and reliance on imperial texts and semi-legendary accounts, archaeology is the only means by which researchers can effectively test claims of pre-state warfare, emergent complexity and early state formation Archaeological context The indigenous community is recognised materially by the Dongson culture (approximately 600 BC–AD 200), widely perceived as providing a foundation for Vietnamese identity First recognised on the basis of excavations at the cemetery and settlement of Dongson, the culture is renowned for its giant ceremonial bronze drums lavishly decorated with ritual scenes and depictions of warriors (Figure 1) (Higham 2004: 58) These objects imply the presence of ranked and complex polities with large populations living in the highly productive agricultural areas of the Bac Bo region (Wheatley 1983; Tessitore 1989; Miksic 2000; Stark 2006) The Dongson culture was spread over a large area of northern Vietnam, and more than 70 sites have been discovered in various environmental settings (Pham 2004: 197) 1012 Research Nam C Kim et al Figure A Dongson drum from Co Loa (tympanum is 73.8cm in diameter with a total height of 53cm) The presence of a unified cultural complex across diverse ecological zones demonstrates considerable interaction between Dongson communities, along with the possibility of either an overarching political structure or perhaps a loose confederation of polities The material evidence suggests the existence of a stratified society (Higham 1996: 132–34), perhaps under the rule of a single state (Bellwood 1992: 125) A Dongson drum was found within the site of Co Loa, a fortified, proto-urban citadel, thought to have been the first political centre or capital of proto-Vietnamese civilisation (Taylor 1983: 20–23) Today, much of the site’s massive earthen ramparts still remain standing, and a national festival is held there annually to commemorate the site and its semi-legendary history Encompassing 600ha, Co Loa’s monumental scale and architecture suggest construction by an early state-level polity Although Vietnamese oral traditions assert the capital site was founded during the third century BC, this claim requires substantiation by scientific investigation 1013 Co Loa: an investigation of Vietnam’s ancient capital The Co Loa fortified site Co Loa is 17km north of present-day Hanoi across the Red River Its surrounding region is highly fertile, comprising 54 per cent of modern agricultural land in northern Vietnam and 70 per cent of all crop production (Higham 1996: 73) Material evidence from within Co Loa’s ramparts and its immediate surroundings indicates a continuous cultural sequence which includes the Phung Nguyen (c 1600–1200 BC), Dong Dau (c 1400–1000 BC), Go Mun (c 1000–600 BC) and Dongson (c 600 BC–AD 200) cultural phases, all of which point to the indigenous emergence of social complexity (Lai 2004; Pham 2004; Nguyen & Vu 2007) The archaeological record confirms the existence of indigenously Vietnamese or proto-Vietnamese communities living continuously within Co Loa’s area from as early as 2000 BC (Lai 2004; Pham 2004) This is supported by archaeological data from over a dozen excavated sites Co Loa’s tightly-knit system of fortifications and earthworks, along with the labour required for construction, demonstrate an integrated system of military defence (Nguyen & Vu 2007: 174–5) These fortifications consist of three massive earthen rampart enclosures or enceintes, river-fed moats and ditches, artificially constructed mounds and towers The ramparts feature exterior ditches, which in ancient times may have been moats connected to a central reservoir located on the site This reservoir was fed by the Hoang River, which in turn was connected to the Red River Ground and aerial surveys show that much of the rampart enclosures still remain standing, though in various states of disrepair, and can be defined as three circuits designated as inner, middle and outer (Figure 2) The inner enclosure is roughly rectangular in shape and measures 1.65km around its perimeter (Nguyen & Vu 2007: 173) Its wall is approximately 5m in height, 6–12m wide on the surface and 20–30m wide at the base (Nguyen & Vu 2007: 173) The inner wall also possesses a number of bastions The middle and outer walls form irregularly-shaped enclosures measuring 6.5km and 8km in circumference, respectively (Nguyen & Vu 2007: 173) It is possible that the irregular shapes stem from the natural topography, with natural hilltops being intentionally connected to form the enclosures In parts, the outer wall still stands 3–4m in height and ranges from 12–20m in width, while the middle wall measures approximately 20m wide and up to 10m in height (Nguyen & Vu 2007: 173–4) Dongson period materials have been found throughout Co Loa In 1970 an informal investigation was conducted by the Vietnamese at a collapsed portion of the outer wall revealing Dongson culture sherds stratified beneath the wall (Nguyen 1970) Moreover, the outstanding Co Loa bronze drum was excavated 500m outside the south-east corner of the inner wall (Pham 1982; Nguyen & Nguyen 1983) The drum, weighing 72kg, is the largest recovered from the region and contained some 200 bronze objects, including 96 ploughshares, six hoes, a chisel and a variety of axes, spearheads, daggers and arrowheads At Cau Vuc, located just outside Co Loa’s southern entrance, a hoard of about 10 000 bronze tanged projectile points was found These artefacts suggest specialised production of tools and weaponry Since Dongson objects are also found throughout the region, Vietnamese researchers have identified a more specific Co Loa culture, distinguishing it from the contemporary Late Dongson of the region (Lai 2004, 2005) 1014 Research Nam C Kim et al Figure A satellite image of the Co Loa earthworks, showing the principal lines of defence In 2004–05, the Vietnam Institute of Archaeology excavated within the inner wall area and several cultural layers were identified The earliest material remains included firing kilns, bricks, stylised ceramic roof tiles and lithic moulds for casting bronze arrowheads A single charcoal sample from a suspected trash pit next to a firing kiln used to produce arrowheads yielded a radiocarbon measurement of 2190+ −35 BP (sample number 05DT.H2f16a; NSFArizona AMS Laboratory number AA75513) (Pham pers comm.) Though overlapping chronologically with the overall Dongson phase, these types of Co Loa artefacts are marked by elite-level or royal characteristics and are found only within the Co Loa site’s enclosures, further reinforcing the notion of centralised production and monopolisation of these materials (Pham 2004; Lai 2005) The middle wall defensive sequence In 2007–08 a new campaign of excavations was undertaken to resolve the sequence of the middle wall at a site adjacent to the North Gate (Figure 2) These excavations cut through the whole width of the rampart and the defensive ditch outside the enclosure to the north 1015 Co Loa: an investigation of Vietnam’s ancient capital Table Five phases of construction of the middle wall enclosure, Co Loa Phase Construction Max height Width Date Clay wall and platform Dumped earth Thick layers of rammed earth Dumped earth Thin layers of rammed earth 1m+ 2m 2.5m 3m 4m 1.8m 17m 24m 24–25m 26m c 4th century BC c 3rd century BC c 1st century AD 15–16th century AD Figure The excavation through the middle wall rampart, looking southwards into the interior of the enclosure The Period wall is visible in the foreground, with the platform and earth structure beyond A 5m wide trench was excavated in 0.1m levels until reaching sterile subsoil (Figure 3) The stratification showed several layers of construction deposits, which could be grouped into three periods and five major phases of construction (Table 1; Figure 4) The Period defence: Phase The earliest structure lay on sterile subsoil and consisted of a clay wall, a clay platform with a structure and ditches associated with them (Figures 3–5) The wall extended in a straight line (in the direction 95◦ E) across our excavation trench into both sections (Figure 5) The 1016 Research Nam C Kim et al Figure The five phases and three periods of construction seen in the west-facing section Phase features date to the fourth century BC or later; Phases 2–4 belong to Period dated to the third century BC or later; Phase is a refurbishment of the fifteenth–sixteenth century AD wall was built with facings of reddish clay and a core of dark brown topsoil The core measured 0.45–0.6m across and the whole width was 1.4–1.8m The wall was visible in section to a height of some 1.1m from the surface of the sterile subsoil The reddish clay had a slumped appearance, suggesting that the wall had been larger when first constructed A ditch measuring approximately 0.5m deep and 1.3–1.5m wide ran along the northern face of the wall On the south side was another ditch with a much gentler slope, reaching a depth of 0.85m and a width of 2.3m Further south was a clay platform that emerged from the west section and ended before reaching the east section Its northern edge ran parallel to the clay wall The platform was surrounded by a ditch On top of the platform was a roughly rectangular earthen structure, approximately 4.4m long and 2m wide and lying approximately 0.4m above the sterile subsoil Taken together these features probably served a military function (Keeley pers comm.), with the platform and its structure perhaps representing a guardhouse, crossbow firing platform or the base of a watchtower This early defensive system appears to be unrelated to the rampart that followed It had a different alignment (95◦ E as opposed to 80–85◦ E) and was thus unlikely to have functioned as a marking-out bank for the rampart’s construction The wall and platform had suffered significant dereliction before the rampart was constructed The dating of the Period features is mainly reliant on radiocarbon (Table 2) The artefacts encountered at this phase and sealed by the rampart were common, non-elite 1017 Co Loa: an investigation of Vietnam’s ancient capital Figure Detail of the Phase clay wall, looking north objects of the Dongson culture, which did not recur in subsequent phases On the floor of the platform’s structure were a number of Dongson potsherds, a fragment of iron slag and a ring of charred earth We secured two charcoal samples from this floor, one from within the fire ring and one with the potsherds and iron slag, which gave dates of 384–114 and 359–54 cal BC (Table 2, CS 138, 139) Ten additional charcoal samples were found with Dongson sherds within the ditch surrounding the platform, sitting on sterile subsoil The radiocarbon age range obtained was 779–56 cal BC (CS 116–8, 120–22, 135–7, 140) One high precision date (CS 136) was 391–209 cal BC The Period defence: Phases 2–4 In the area where it was examined, the main rampart ran in a generally east-west direction (80–85◦ E) and in its final surviving form was 4.3m in height and 26m wide at the base There were four phases of construction, all of earth Although it is difficult to determine how much time separated these major construction events, our interpretation is that most of the rampart (Phases 2–4) was constructed continuously within a relatively rapid time-span This is because there were no layers of natural deposition or erosion which would be expected had there been large temporal gaps between construction phases (Keeley pers comm.) However, 1018 Nam C Kim et al Sample no (From the structure floor) 07TTH1CS138 07TTH1CS139 (From the ditch) 07TTH1CS116 07TTH1CS117 07TTH1CS118 07TTH1CS120 07TTH1CS121 07TTH1CS122 07TTH1CS135 07TTH1CS136∗ 07TTH1CS137 07TTH1CS140 ∗ Radiocarbon age Calibrated (1σ ) Calibrated (2σ ) 2150+ −43 BP 2186+ −44 BP 351–111 BC 358–191 BC 359–54 BC 384–114 BC 2365+ −52 BP 2446+ −83 BP 2392+ −51 BP 2327+ −52 BP 2196+ −31 BP 2205+ −31 BP 2154+ −43 BP 2251+ −22 BP 2192+ −33 BP 2282+ −30 BP 520–387 BC 749–410 BC 704–397 BC 506–236 BC 356–201 BC 358–204 BC 353–113 BC 385–234 BC 356–199 BC 397–259 BC 751–260 BC 779–398 BC 752–387 BC 727–207 BC 369–181 BC 377–197 BC 360–56 BC 391–209 BC 370–174 BC 402–211 BC Sample randomly selected for high-precision analysis (multiple run) and weighted average shown here based on artefacts, there was a final Phase assigned to a refurbishment in a later period (see below) Construction of the main rampart began with a foundation of spoil dug out from the ditch (Keeley pers comm.) Dumps of soil, sand and clay were then deposited in layers to form a mounded, bank-like feature (Phase 2) A depression was clearly visible along the top of the Phase bank, suggesting its use as a pathway for walking, pulling carts and transporting material This path was no doubt brought into use in the construction of the next level of rampart, Phase 3, which was of rammed or stamped earth This left the rampart as a whole higher and stronger, and with a flat top Phase involved another layer of dumped earth, further expanding the size of the rampart Most of the artefacts, consisting of ceramic roof tiles and stones, were recovered from within the Phase layers (approximately 1m below the current rampart surface) where they sloped towards both north and south (Figure 6), although there were far more artefacts on the south side than the north Roof tiles and stones were also found in a ditch towards the south, which may represent the inner edge of the rampart In the north, the artefacts were also found in the defensive ditch, where they had probably been displaced from the top of the rampart The presence of the roof tiles and stones, which are part of the Co Loa elite material culture, is still open to interpretation One possibility is that there was a roofed structure erected along the top of the rampart to protect defenders against rain or missile attack Another possibility is the tiles were placed there intentionally to protect the rampart core from erosion by tropical rain Whatever the case, the debris of tiles and stones seems to exist along the same stratigraphic layer throughout both the middle and outer walls, as indicated by our surveys at collapsed portions of the ramparts 1019 Research Table Radiocarbon ages for charcoal found with Dongson potsherds on the floor of the platform structure and in the ditch of the platform mound Period 1, Phase (calibrated to Reimer et al 2004) Co Loa: an investigation of Vietnam’s ancient capital Figure Detail of the Phase level with dumped roof tiles The roof tiles (Figure 7) are diagnostically similar to those found within the inner wall area in 2005, and some of the stones appear to be pre-forms or discarded blanks of casting moulds for bronze crossbow bolts, also similar to examples found within the inner wall area Eleven radiocarbon dates suggest that the Period rampart was under construction in the third century BC and remained in active use until the end of the millennium (Table 3; see Dating below) The defensive ditch In order to study the defensive ditch, a × 46m trench was opened at the base of the rampart and excavated in 0.1m levels until sterile soil was reached Unfortunately, the ditch excavation was interrupted by a collapse within the trench due to heavy flooding Hence, our current understanding of the ditch is cursory and future research will be necessary for a more thorough analysis (Figure 8) The visible stratigraphy consisted of 17 depositional layers, though whether deposition was natural or artificial is currently difficult to assess Based on the stratigraphic evidence and the location of cultural materials, the ditch was approximately 10m wide at the top and V-shaped in profile The ditch was dry when originally constructed, but may have functioned as both a dry ditch and a water-filled moat after construction, as indicated by the presence of alluvial 1020 Nam C Kim et al Table Radiocarbon ages for charcoal found within the Middle Period rampart wall Period 2, Phases 2–4 (calibrated to Reimer et al 2004) 07TTH1CS110 ∗ 07TTH1CS88 07TTH1CS111 ∗ 07TTH1CS88 07TTH1CS76 07TTH1CS41 07TTH1CS69 07TTH1CS70 07TTH1CS9 07TTH1CS63 07TTH1CS50 ∗ Phase Radiocarbon age Calibrated (1σ ) 3 3 4 4 4 2264+ −39 BP 2253+ −39 BP 2234+ −41 BP 2184+ −34 BP 2187+ −33 BP 2184+ −43 BP 2170+ −33 BP 2139+ −33 BP 2136+ −44 BP 2116+ −43 BP 2093+ −43 BP 392–233 BC 389–215 BC 380–210 BC 356–192 BC 356–196 BC 357–184 BC 353–173 BC 344–111 BC 347–93 BC 198–59 BC 168–54 BC Sample randomly selected for high-precision analysis (multiple run) Figure A roof tile in situ 1021 Calibrated (2σ ) 399–206 BC 396–204 BC 389–203 BC 375–165 BC 370–170 BC 382–154 BC 363–113 BC 353–54 BC 357–46 BC 353–4 BC 345 BC–AD Research Sample no Co Loa: an investigation of Vietnam’s ancient capital Table Radiocarbon ages for charcoal found within the ditch of the Middle Period rampart wall (calibrated to Reimer et al 2004) Sample no 07TTH1CS108 07TTH1CS127 Radiocarbon age Calibrated (1σ ) Calibrated (2σ ) 2030+ −52 BP 2050+ −110 BP 174 BC–AD 72 200 BC–AD 67 103 BC–AD 48 380 BC–AD 208 Figure Profile of the Period defensive ditch layers consisting of soil, clay, silt, sand and gravel (Green pers comm.; Junker pers comm.) The ditch’s shape and function undoubtedly changed over time, and changes may have also occurred seasonally We suspect the ditch served a defensive function, an interpretation bolstered by both the military function of the rampart and by the V-shape of the ditch A range of artefacts was recovered, from Co Loa culture at the bottom to historic Le Dynasty materials within the upper strata The lowest cultural layers had Co Loa roof tiles and stones, similar to the ones found within the rampart excavation Two wood charcoal samples were recovered from the ditch excavation (Table 4) The first sample was recovered in association with Co Loa roof tiles found at the lowest cultural level at approximately 4m below the surface, and gave a date range of 200 cal BC–cal AD 67 The second sample, recovered through an augur core, was just above the suspected sterile soil at 4.3m below the surface at the bottom of the ditch’s V-shape The radiocarbon age is 103 cal BC–cal AD 48 While these radiocarbon dates not reveal when the ditch was originally dug, they offer a rough chronological sense of when the roof tiles may have begun eroding from the rampart into the ditch The Period defence The Phase construction expanded the rampart to nearly its final surviving dimensions (Figure 4) The mode of construction employed multiple, thin layers of stamped earth Below the uppermost surface, there are stepped tiers on both exterior faces The first tier of the southern interior side may have served as a parapet of some kind, allowing defenders to move unseen along the ledge while under protection from potential projectile attack 1022 The dimensions for the northern exterior face are different and may appear stepped due to natural erosion and collapse Artefacts of the Vietnamese Le Dynasty (c fifteenth–sixteenth centuries) were found within these upper layers, suggesting a refurbishment in the historical period The Phase construction layers appeared to be thin and uniform, resembling the hang-tu method of Chinese construction (Keeley pers comm.) While Le Dynasty materials were found, it is possible that refurbishment or amplification efforts began during the Nam Viet (Nan Yue) period or the Chinese colonial period (beginning with the Han) Dating Most of the date ranges provided by the 12 radiocarbon determinations for Period included the period 400–350 BC, with four of these samples yielding somewhat earlier dates (Table 2) Given the likelihood of residual and intrusive material, our reading of these dates is that they probably represent ground cleared for construction in the fourth century BC For Period (Table 3), one radiocarbon date was secured from wood charcoal found on the interface between the Phase layer of clumped soil and the sterile soil beneath it of 399–206 cal BC (CS110) Dates from Phase layers group between 396–170 cal BC (CS 76, 88 and 111), with two high precision dates on CS 88 giving 396–165 cal BC Six radiocarbon dates were secured from wood charcoal samples found with the roof tiles in Phase layers These cover a range from 382 BC–cal AD The two radiocarbon dates from the ditch covered the span 103 BC–cal AD 67 (Table 4) Again making allowances for residual and intrusive charcoal, we suspect the construction of the Period rampart may have commenced sometime in or after the third century BC, and that it continued in existence until its disuse, probably in the early years of the first millennium AD Period (Phase 5), dated by pottery, would appear to be a refurbishment of the fifteenth– sixteenth centuries AD, although amplification efforts may have commenced much earlier Overall, the combination of artefacts, contrasting building techniques and radiocarbon dates currently suggest that the majority of the middle wall rampart was constructed by a local and indigenous society prior to the first century AD solidification of Han colonial control Given the timing of construction, it is possible that turmoil in China during the Warring States period may have played some role in motivating fortification That said, more research is needed to further refine this chronology and improve our overall understanding of the cultural history For instance, while the material evidence lends support to the possible existence of the Au Lac polity, it does not prove it unequivocally Moreover, it is possible that the Nan Yue polity’s attacks and eventual overthrow (c early second century BC) of the Au Lac, as described in textual accounts, may have been responsible for either precipitating rampart construction or amplifying it after takeover The stamped earth technique visible within Phase (Period 2) bears some resemblance to stamped earth techniques (hang-tu) seen in parts of ancient China The hang-tu method was often used to construct walls and foundations for buildings at Longshan (c 3000– 1800 BC) and Shang culture (c 1600–1046 BC) sites (Chang 1978, 1980) However, stamped earth layers at Han Chinese sites tended to be thinly stratified and uniform in thickness, approximately 120–140mm, while the stamped earth layers of Phase were much 1023 Research Nam C Kim et al Co Loa: an investigation of Vietnam’s ancient capital thicker, cruder, and lacked uniformity (Keeley pers comm.) The material record of various civilisations shows use of a stamped earth technique, as indicated by the archaeological and ethnohistoric records of southern Mesopotamian city-states, the Inka of Peru and the Yoruba states of Africa (Trigger 2003: 566) Consequently, we not believe that this type of construction necessarily indicates that the wall was a Chinese construction, although this does not exclude the possibility of Chinese influence or emulation Interpretation and discussion The earliest features of Period predate the rampart constructed in Period What remains unclear is how much time separated the two The scale of construction of the early features is more modest than what was to follow, implying a lower availability of labour and technical capacity, with perhaps a lesser scale of perceived threat Of course, our understanding of this smaller set of defensive works is far from complete, as we have only unearthed a portion of it Future investigations will need to examine these features further, and should investigate whether or not similar features exist elsewhere at Co Loa For the time being, though, the smaller fortifications represent the earliest that have been archaeologically detected anywhere within Vietnam They also provide support for the occurrence of intra-regional competition and warfare among smaller-scale, non-state societies of the Red River valley during the mid first millennium BC, suggesting the possibility that conflict contributed to political consolidation (Carneiro pers comm.) Throughout the region, there is ample archaeological evidence of warfare in the preDongson and Dongson eras, including stone and bronze weapons (Pham 2004: 199) and iconographic depictions of warriors and war captives on the famous Dongson bronze drums themselves (Higham 2004: 58) The evidence underscores a political landscape of multiple competing polities, possibly chiefdom-level, during the mid first millennium BC wherein elites used a combination of strategies to obtain economic wealth and entrench political power Military power and coercive force, especially after the introduction of bronze weapons, thus constituted one means for consolidating power within a wider causal package Because of its strategic location near the Red River, elites living on the Red River plain were also probably able to accumulate considerable wealth and political currency through the exchange of bronze prestige goods within an interregional trade network that connected southern China, northern Vietnam and parts of Southeast Asia through riverine systems (Allard 1999; Stark 2006: 414; Kim 2009) These factors, combined with the agricultural productivity of the area, afforded tremendous opportunities for certain societal segments to generate politically exploitable surpluses (Junker pers comm.) In addition, materialised ideology may have been used as part of power strategies that reconfigured sociopolitical systems (DeMarrais et al 1996) The iconography of Dongson drums encapsulated the existing sociopolitical order by depicting and codifying the exclusive rituals of Dongson elites This provides a context for the massive scale of Co Loa’s awe-inspiring ramparts, which would have also served an ideological function, displaying the level of authority present and its ability to mobilise labour and harness resources Overall, Co Loa’s ramparts suggest the presence of significant centralised authority which probably required military force 1024 to obtain and keep As such, reconstructing the site’s history can thus inform ongoing discussions regarding emergent complexity and the origins of Vietnamese civilisation In Southeast Asia, large enclosed and moated settlements dating to the first millennium BC have also been found in Thailand and Cambodia, although the function and chronology are not altogether clear for many of them (Moore 1992; Higham 2002) Several such sites are distributed in the Mun River valley, such as Non Dia, Noen U-Loke and Non Muang Kao (Higham 2002: 193–204) and also within the Chi Valley, such as Non Chai and Ban Chiang Hian (Higham 1996: 214–5) Elsewhere, numerous red soil circular earthwork sites of south-east Cambodia and southern Vietnam, such as Krek 52/62, also possessed embankments and moats (Albrecht et al 2000: 43) Another example, the late prehistoric site of Khao Sam Kaeo (c fourth–second century BC) in the upper Thai-Malay Peninsula, possesses a fortification system of embankments and ditches that bear some resemblance to those found at Co Loa (Bellina-Pryce & Silapanth 2006) In the case of Khao Sam Kaeo, the complex earthwork system, combined with specialised craft production and exchange, also hint at the presence of a significant early polity (Bellina-Pryce & Silapanth 2006: 286) Conclusion Undoubtedly, much more research and data are needed to augment our understanding of the prehistory of Co Loa and the Red River plain For instance, a consideration of individual strategies and behaviour, and their effects on social structure, would require data on households and settlements, which are currently lacking for all of Southeast Asia (Bellwood & Glover 2004: 11) Examining Dongson land use and agricultural development strategies could help identify incipient urbanism (Kealhofer & Grave 2008: 201) A regional site survey could also help determine whether or not a settlement size hierarchy is discernible However, the investigations at Co Loa, compared with the currently available evidence of written sources and archaeology, indicate that a local and indigenous state-like polity had emerged in the Bac Bo region during the Dongson period and prior to Han colonisation The sheer size and scale of Co Loa and its monumental fortifications strongly suggest a high degree of political centralisation was necessary to plan construction and mobilise the requisite resources Prior to Co Loa, the region’s communities were likely already accustomed to construct communal works, even if small-scale (e.g ditches, dikes, walkways) (Larew 2003: 40–41) However, nothing on the scale of Co Loa had ever been attempted before, and the corv´ee labour necessary for its construction must have required a strong military force and significant centralised, state-like control (Larew 2003: 41) Furthermore, the labour requirements strongly imply a high population density A large population would have been sustainable due to the significant rice paddy agricultural potential of the Red River valley region, which today produces several crop yields annually Despite the evidence of continuous, in situ cultural development during the late Neolithic and early Metal Age, the Red River plain’s communities did not live in a vacuum It is highly likely that two-way influence between Bac Bo societies and their Chinese counterparts occurred before and during the Dongson period (Allard 1999: 83) While emergent local complexity is not directly attributable to Han colonisation, first-millennium BC interregional contact and exchange between Dongson communities and stratified societies of 1025 Research Nam C Kim et al Co Loa: an investigation of Vietnam’s ancient capital neighbouring regions (such as contemporaneous Dian societies of China’s Yunnan area, Sa Huynh communities of central Vietnam and upland societies to the west and southwest in parts of Laos, Cambodia and Thailand) almost certainly contributed to a process of reciprocal complexity (Kim 2009) Given the likelihood that the bulk of the Co Loa rampart was constructed during the third century BC, at the tail end of the Warring States period and during the period of Chinese unification under Qin imperial power, it is conceivable that an external foreign threat may have also influenced political trends in the Bac Bo region, in addition to local, intra-regional competition (Higham 2002: 170) Further, there are signs of Chinese influence or emulation in some of the cultural elements found at Co Loa, including the stamped earth technique, bronze crossbow bolts, and roof tiles Elsewhere, Larew (2003) examines the possible transfer of military technologies from Warring States China to the Red River plain and Co Loa Future research would be well served by continuing to explore the impact of the powerful Chinese states to the north on Bac Bo societies, even before the latter came to be fully colonised by the Han, as well as interaction with neighbouring societies throughout Southeast Asia Acknowledgements The Co Loa defences project was co-directed by Dr Nam Kim and Dr Lai Van Toi and involved additional archaeologists from both the University of Illinois at Chicago (Drs Laura Junker, Lawrence Keeley and Debra Green), the Vietnam Institute of Archaeology (Dr Trinh Hoang Hiep and Mr Nguyen Dang Cuong), and the Conservation Center for the Co Loa and Hanoi Citadels (Ms Nguyen Thi Thuy) We would like to thank the American Council of Learned Societies, the Henry Luce Foundation, the American Philosophical Society, the George Franklin Dales Foundation and the National Science Foundation (Award No 0915410) for funding that made this research possible Radiocarbon dates were analysed at the NSF-AMS Laboratory at the University of Arizona We would also like to thank the Vietnam Institute of Archaeology, the Conservation Center for the Co Loa and Hanoi Citadels, the Vietnam Academy of Social Sciences and the Vietnamese Ministry of Sports, Culture, and Tourism The image of the Co Loa bronze drum was provided courtesy of the Vietnam Social Science Publishing House The authors are especially grateful for constructive comments and questions provided for an earlier version of this paper by Francis Allard, Ian Glover, Charles Higham, Laura Junker, Lawrence Keeley and TzeHuey Chiou-Peng as well as Martin Carver and the editorial staff of Antiquity Any errors in the paper are the responsibility of the authors Finally, we wish to thank the people of the Co Loa Commune References ALBRECHT, G., M.N HAIDLE, C SIVLENG, H.L HONG, H SOPHADY, H THAN, M SOMEAPHYVATH, S KADA, S SOPHAL, T CHANTHOURN & V LAYCHOUR 2000 Circular earthwork Krek 52/62: recent research on the prehistory of Cambodia Asian Perspectives 39(1–2): 20–46 ALLARD, F 1999 The archaeology of Dian: trends and tradition Antiquity 73: 77–85 BELLINA-PRYCE, B & P SILAPANTH 2006 Weaving cultural identities on trans-Asiatic networks: Upper Thai-Malay Peninsula – an early socio-political landscape Bulletin de l’Ecole francaise d’Extreme-Orient 93: 257–83 BELLWOOD, P 1992 Southeast Asia before history, in N Tarling (ed.) The Cambridge history of Southeast Asia: 55–136 Cambridge: Cambridge University Press BELLWOOD, P & I GLOVER 2004 Southeast Asia: foundations for an archaeological history, in I Glover & P Bellwood (ed.) Southeast Asia: from prehistory to history: 4–20 New York: Routledge CHANG, K.C 1978 The archaeology of China New Haven (CT): Yale University Press – 1980 Shang civilization New Haven (CT): Yale University Press DEMARRAIS, E., L.J CASTILLO & T EARLE 1996 Ideology, materialization and power strategies Current Anthropology 37(1): 15–31 1026 HIGHAM, C 1989 The archaeology of mainland Southeast Asia Cambridge: Cambridge University Press – 1996 The Bronze Age of Southeast Asia Cambridge: Cambridge University Press – 2002 Early cultures of mainland Southeast Asia Bangkok: River Books – 2004 Mainland Southeast Asia from the Neolithic to the Iron Age, in I Glover & P Bellwood (ed.) Southeast Asia: from prehistory to history: 41–67 New York: Routledge KEALHOFER, L & P GRAVE 2008 Land use, political complexity and urbanism in mainland Southeast Asia American Antiquity 73(2): 200–225 KIM, N 2009 Terrain, interregional interaction and reciprocal complexity in Metal Age northern Vietnam and southern China Unpublished manuscript LAI, V.T 2004 Co Loa archaeology in past years and future prospect, in T Nguyen (ed.) Mot The Ky Khao Co Hoc Viet Nam: Tap I: 670–71 Ha Noi: Vien Khoa Hoc Xa Hoi Viet Nam, Vien Khao Co, Nha Xuat Ban Khoa Hoc Xa Hoi – 2005 Tu lieu Vien Khao co hoc, dot khai quat nam 2004–2005 Ha Noi: Vien Khao Co Hoc LAREW, M 2003 Thuc Phan, Cao Tong, and the transfer of military technology in third century BC Viet Nam East Asian Science, Technology, and Medicine 21: 12–47 MIKSIC, J 2000 Heterogenetic cities in pre-modern Southeast Asia World Archaeology 32(1): 106–120 MOORE, E 1992 Water enclosed sites: links between Ban Takhong, northeast Thailand and Cambodia, in J Rigg (ed.) The gift of water: water management, cosmology and the state in South East Asia: 26–46 London: School of Oriental and African Studies NGUYEN, D.T 1970 Phan tich bao tu phan hoa Co Loa Khao Co Hoc 12(7–8): 145–6 NGUYEN, G.H & V.H NGUYEN 1983 Nhom Do Dong Moi Phat Hien O Co Loa (Ha Noi) Khao Co Hoc 47: 21–32 NGUYEN, Q.N & V.Q VU 2007 Dia Chi Co Loa Hanoi: Nha Xuat Ban Ha Noi O’HARROW, S 1979 From Co-loa to the Trung sisters’ revolt: Vietnam as the Chinese found it Asian Perspectives 22(2): 140–63 PHAM, H.T 1982 Phat Hien Co Loa 1982 Hanoi: Vietnam Institute of Archaeology PHAM, M.H 2004 Northern Vietnam from the Neolithic to the Han Period – Part 2: the Metal Age in the north of Vietnam, in I Glover & P Bellwood (ed.) Southeast Asia: from prehistory to history: 189–201 New York: Routledge REIMER, P.J., M.G.L BAILLIE, E BARD, A BAYLISS, J.W BECK, C.J.H BERTRAND, P.G BLACKWELL, C.E BUCK, G.S BURR, K.B CUTLER, P.E DAMON, R.L EDWARDS, R.G FAIRBANKS, M FRIEDRICH, T.P GUILDERSON, A.G HOGG, K.A HUGHEN, B KROMER, G MCCORMACK, S MANNING, C.B RAMSEY, R.W REIMER, S REMMELE, J.R SOUTHON, M STUIVER, S TALAMO, F.W TAYLOR, J VAN DER PLICHT, C.E WEYHENMEYER 2004 IntCal04 terrestrial radiocarbon age calibration, 0–26 cal kyr BP Radiocarbon 46: 1029–58 STARK, M 2006 Early mainland Southeast Asian landscapes in the first millennium AD Annual Review of Anthropology 35: 407–432 TAYLOR, K 1983 The birth of Vietnam Berkeley (CA): University of California Press TESSITORE, J 1989 View from the East Mountain: an examination of the relationship between the Dong Son and Lake Tien civilizations in the first millennium BC Asian Perspectives 28(1): 31–44 TRIGGER, B 2003 Understanding early civilizations Cambridge: Cambridge University Press WHEATLEY, P 1983 Nagara and commandery: origins of Southeast Asian urban traditions (University of Chicago, Department of Geography, Research Paper 207–8) Chicago (IL): Department of Geography, University of Chicago 1027 Research Nam C Kim et al  .. .Co Loa: an investigation of Vietnam’s ancient capital Historical context The conventional reconstruction of the protohistory of the northern Vietnam region during the second half of the... 1013 Co Loa: an investigation of Vietnam’s ancient capital The Co Loa fortified site Co Loa is 17km north of present-day Hanoi across the Red River Its surrounding region is highly fertile, comprising... north 1015 Co Loa: an investigation of Vietnam’s ancient capital Table Five phases of construction of the middle wall enclosure, Co Loa Phase Construction Max height Width Date Clay wall and platform