This document discusses water management systems in Funan and Chenla in Southeast Asia from the 1st millennium CE. It provides background on the regions' topography, climate and debates around Indian influence. The author argues that while India likely played a role in introducing new technologies and religions, Funan and Chenla's systems drew on local environments and assimilated foreign ideas rather than purely replicating Indian models. Case studies of sites like Angkor Borei, Noen U-Loke, Wat Phu and Sambor Prei Kuk are used to analyze water management approaches in each kingdom within their historical contexts.
1. Water management systems from
Southeast Asia: Funan and Chenla
Brice Detune
Dissertation submitted in partial fulfilment of the requirements for the
degree of MA in Archaeology and Heritage of Asia of University
College London in 2014
UCL INSTITUTE OF ARCHAEOLOGY
2. 2
Abstract
In the following dissertation I address the water management systems from
Funan and Chenla. I try to challenge their potential continuity and the influence
India could have had on them. I suggest that those from Funan were based on the
natural environment while those from Chenla are more independent. Furthermore
the water management systems from Chenla show a will to store water that could
probably be linked to irrigated agriculture. India certainly had a role in it since the
introduction of new water management systems seems to be correlated with the
arrival of Hindu temples that would suggest monastic landlordism. Furthermore it
is possible that oryza sativa indica would have been introduced at the same time.
However Chenla’s water management systems still draw on Funan’s previous
technologies and display instead an assimilation of Indian technologies and not
pure Indian technologies
Acknowledgements
First of all I wish to thank Pr. Julia Shaw, my advisor, for offering me guidance
on this dissertation and particularly for helping me focus on a topic. Dr. Cristina
Castillo has been much helpful as well in my understanding of the issues
surrounding rice that are incredibly more complex than what I have ever thought. I
also owe much to Pr. Miriam Stark and Pr. Peter Sharrock for answering my
questions. And finally I would like to thank UCL’s Institute of Archaeology for giving
me the opportunity to write the following dissertation.
3. 3
Table of Contents
List of Figures ..................................................................................................................................... 5
Introduction ....................................................................................................................................... 6
I. General background information............................................................................................... 7
A. Topography and main features.............................................................................................. 7
B. Climate ................................................................................................................................... 7
II. Theoretical framework............................................................................................................... 9
A. The Indianization.................................................................................................................... 9
1. The concept of Indianization.............................................................................................. 9
2. How did Indianization happen and what did it involve?.................................................. 11
Conclusion................................................................................................................................ 15
B. Funan.................................................................................................................................... 16
1. A polity known thanks to China ....................................................................................... 16
2. The debates surrounding Funan ...................................................................................... 19
Conclusion................................................................................................................................ 19
C. Chenla .................................................................................................................................. 20
1. Chenla, Funan and China.................................................................................................. 20
2. Creation of the kingdom of Chenla.................................................................................. 21
3. Water Chenla and Land Chenla........................................................................................ 22
Conclusion................................................................................................................................ 23
Chapter conclusion ...................................................................................................................... 24
III. The use of water .................................................................................................................. 25
A. Agriculture: rice.................................................................................................................... 25
1. Growing rice..................................................................................................................... 26
2. A short history of the domestication of rice and its dispersal......................................... 28
Conclusion................................................................................................................................ 32
B. Water transportation........................................................................................................... 33
C. Water, religion and agriculture............................................................................................ 34
Chapter conclusion ...................................................................................................................... 36
IV. Case study ............................................................................................................................ 37
A. Methodology........................................................................................................................ 37
4. 4
B. Water management systems in Funan ................................................................................ 38
1. Angkor Borei..................................................................................................................... 38
2. Noen U-Loke..................................................................................................................... 46
Conclusion................................................................................................................................ 49
C. Water management systems in Chenla ............................................................................... 50
1. Wat Phu............................................................................................................................ 50
2. Sambor Prei Kuk............................................................................................................... 56
Conclusion................................................................................................................................ 60
D. Water management systems in India .................................................................................. 60
Chapter conclusion ...................................................................................................................... 63
Conclusion: Funan and Chenla, a water management perspective................................................. 64
References cited............................................................................................................................... 67
5. 5
List of Figures
Figure I.1: Map of the physical setting of Southeast Asia.................................................................. 8
Figure II.1: Statues of Vishnu and Ganesh found during excavations in the Mekong Delta............ 10
Figure II.2: Some carnelian and agate beads from Khao Sam Kaeo................................................. 11
Figure II.3: A suggested map of the trade routes in Asia around ca. 600 CE................................... 12
Figure II.4: Another possible map of the trade routes in Asia in the mid-1st
millennium CE. ......... 13
Figure II.5: Rock crystal lion found in Phu Khao Thong.................................................................... 14
Figure II.6: Carnelian lion found in Ban Don Ta Phet....................................................................... 14
Figure II.7: Conch shell seal found in Phu Khao Thong. ................................................................... 14
Figure III.1: Map of the zone occupied by wild rice during the Last Glacial (20,000 BP) and its
expansion in the Early Holocene (9,000 BP). ................................................................................... 25
Figure III.2: Table of the burial phases and samples used for δ13
C analyses in Thailand ................ 30
Figure III.3: Results from the analyses in the Upper Mun River Valley............................................ 31
Figure III.4: Results from the analyses in Ban Chiang. ..................................................................... 31
Figure IV.2: Archaeological map of Angkor Borei after the 1995-1996 excavation......................... 39
Figure IV.3: Canals recognized after aerial observations in the middle of the 20th century. ......... 44
Figure IV.4: Map of the canals around Angkor Borei showing the palaeochannels........................ 45
Figure IV.5: Aerial photography of Noen U-Loke............................................................................. 47
Figure IV.6: Drawing of the mound of Noen U-Loke and of its moats............................................. 48
Figure IV.7: Aerial photography of Wat Phu.................................................................................... 52
Figure IV.8: Archaeological map of the ancient city of Wat Phu ..................................................... 55
Figure IV.9: The complex of Sambor Prei Kuk.................................................................................. 58
Figure IV.10: The city area of Sambor Prei Kuk................................................................................ 59
Figure IV.11: Estimated annual runoff in India and name of the regions........................................ 62
6. 6
Introduction
Funan and Chenla, although being the precursors of the Angkorian Empire are
unfortunately poorly known and hardly studied. However the mastery in water
technology shown by the City of Angkor certainly had a past, hence it is important
to understand what was the level of technical knowledge of these past polities in
hydraulic and more simply what means they possessed. Moreover the transition
from Funan to Chenla remains unclear and major changes may have happened.
Another common issue when dealing with polities in Southeast Asia is the
impact of the Indian influence. Since the subcontinent already had developed
water management systems it is not possible to overlook the idea that it might
have had a role in the development of those systems in Funan and Chenla.
Hence I will try to answer the following questions:
- Is there continuity between the water management systems from Funan
and those from Chenla?
- Do we see any evidence for an Indian influence?
To achieve this goal it seemed logical to focus on case studies since
archaeology should be first preoccupied by the evidence. However, before these
case studies I will present some general background information and some
theoretical issues that have to be acknowledged in order to interpret the results.
Dealing primarily with water I will also present its possible uses prior to studying
the evidence.
Finally I will conclude with a water management perspective on Funan and
Chenla.
7. 7
I. General background information
A. Topography and main features
The area we are going to study is located in Peninsular Southeast Asia,
focusing on Cambodia, Vietnam, Laos and Thailand. The environment is very
diverse ranging from mountains to the North, in Northeast Thailand and Laos, to
flat low-lying areas such as the Mekong Delta (Fig. I.1). The difference between
both is very important since it creates a real separation between these zones and
also offers natural borders. The types of habitats developed obviously also depend
on the slopes that are increasing as we go North towards the border with China.
A great part of Peninsular Southeast Asia is bordered by the sea and many
people still permanently live on boats. Furthermore there are several rivers, the
most important of them being the Mekong that is used as a major axis nowadays.
B. Climate
Southeast Asia is characterized by two seasons, a dry season and a wet
season, due to the monsoon. The dry season lasts from November to April and
water is scarce. In the wet season it is exactly the opposite as it rains nearly every
day and the lowlands are inundated usually from July to November. The
inundations can reach more than 3 meters Above Mean Sea Level (AMSL). This
particularly concerns the Mekong Delta area where the only inhabited places at
that time of the year are on mounds and the only mean of transportation is boat.
There is evidence that the Sea Level has changed over time however it seems
it does not have much impact in the Mekong Delta area for the periods that
interest us.
8. 8
Figure I.1: Map of the physical setting of Southeast Asia. We clearly see the vast majority of the land is made of low-
lying plains. From Bellwood 1999: Map 2.1
9. 9
II. Theoretical framework
To be able to understand fully the topic of this dissertation, it is important to
understand a certain number of concepts I will review here, such as the
Indianization or what were Funan and Chenla.
A. The Indianization
The concept of Indianization is absolutely central whenever one is addressing
the Early History of Southeast Asia and this dissertation is no exception to that
rule. We will see in this section what lies beyond the concept of Indianization that
evolved a lot since it has first been hypothesized and how it may have happened.
1. The concept of Indianization
From the beginning of the 20th century and up until recently, the most common
point of view held among scholars was that India eventually brought civilization to
Southeast Asia (Coedès 1975; Higham 2001; 2013; Stark & Allen 1998). In fact,
from the 16th century onwards, Southeast Asia was considered under-evolved and
the Europeans could not believe that the locals could have been able to build the
architectural realizations they saw (Higham 2001: 2; 2013: 1-2). Around the
beginning of the 20th century an Indian origin had been suggested, mainly because
of the similarities with the buildings from India and Sanskrit inscriptions found in
the area (Higham 2001). Archaeology, that developed much later in Southeast
Asia because of political instabilities mainly (Stark 1998), seemed to back the idea
in the beginning, recovering artefacts that were Indian-like.
More recently however, the concept has been challenged. Closer looks to the
evidence have revealed anomalies pointing towards an appropriation of the Indian
codes by the indigenous. The temples of the Dieng plateau in central Java for
10. 10
instance are not exact transcriptions of the Indian codes; they exhibit some
typically local elements that are not to be found anywhere else (Romain 2011).
Moreover those free-standing temples have been built at the same time as the
free-standing ones in India suggesting a strong connection between the two
countries that allowed a quick circulation of ideas. The same can be said about
sculpture and Stargardt (2003: 110-11) has shown the style used for the
sculptures of Vishnu in the Kra Isthmus, although containing some influences from
the Gupta, the post-Gupta and the Pallava, have some unique features. (Fig. II.1)
Figure II.1: Statues of Vishnu and Ganesh found during excavations in the Mekong Delta. From Malleret 1959: Plate
LXXXV
Even the artefacts that were thought to shed light on the early phases of
Indianization, from the 4th century BC onwards are problematic. Beads, for
instance, considered as typically Indian and found all over the area from Thailand
to the Philippines (Bellina 2003; Dussubieux & Gratuze 2003; Glover & Bellina
2011; Gupta 2003;) appear to have been crafted in Southeast Asia, as in Khao
Sam Kaeo or Ban Don Ta Phet (Glover & Bellina 2011) (Fig. II.2). Moreover some
lapidary techniques used in these sites are barely seen in India (Glover & Bellina
2011).
11. 11
Figure II.2: Some carnelian and agate beads from Khao Sam Kaeo. From Glover & Bellina 2011: Figure 2.12
Hence, for some scholars, too much importance granted to art and the
presence of Sanskrit gave the impression Southeast Asia was Indian (Stark 1998).
In fact it seems Southeast Asia had many exchanges with India before the
Christian Era and that the artefacts recovered are a sign of growing trade and not
of the presence of Indians. Later the only real elements used to suppose Indian
influence are all related to religion. The concept of Indianization has then evolved
to be defined as a sphere of common conceptions and shared technological
knowledge and not merely the idea the India brought civilization to Southeast Asia
(Gupta 2003; Stark 1998). The creation of this sphere would have begun as soon
as the 4th century BC from archaeological evidence (Gupta 2003).
2. How did Indianization happen and what did it involve?
There are four theories to explain the diffusion of Indian cultural traits in
Southeast Asia:
- The Kshatriya theory: they were brought through war
- The Brahmin theory: they were brought by Brahmins trying to convert them
- The Vaishya theory: they were brought by merchants
- A mix of the above theories
12. 12
As said previously, from the evidence we have there have been regular
contacts through trade (from the fourth century BC onwards (Guy 2011) (Fig. II.3 &
II.4). Those were materialized partly through beads and vessels.
There is little evidence to support the Kshatriya theory as not traces of war
have been found (Guy 2011).
It has been noticed by several authors that Hinduism prevents people from
going across the borders for the foreigners are not pure; therefore it is unlikely
they represented the first contacts with Southeast Asia (Coedès 1975; De
Casparis et al. 1992).
Figure II.3: A suggested map of the trade routes in Asia around ca. 600 CE. It is a very global one including the inland
Silk Road going through the Taklamakan desert; hence it gives no indication about the relative importance of each
possible path. Nevertheless it shows the global connections that already existed around the mid-1st millennium CE.
However the relation between the western coast and the eastern coast of India at the southern-most point is
arguable. From Khoo 2003
13. 13
Figure II.4: Another possible map of the trade routes in Asia in the mid-1st millennium CE. That one privileges a trans-
isthmian way. In fact, according to Stargardt (2003), it is highly probable that both routes, the trans-isthmian and the
circum-peninsular ones, have been used and that they were more or less important depending on the time in History.
Furthermore, this map suggests a possible link with Gandhara at the western end of the route. As demonstrated by
Gupta (2003), this is highly improbable; otherwise the Hindu sculptures we find in Southeast Asia would have
developed earlier. From Jacques 2007
Hence the process would have started though trade. However, these traded
objects have more than a material aspect; they bring along some cultural ideas
(Gupta 2003; Stargardt 2003). For instance a carnelian lion, more than a simple
object, can be perceived as a metaphor of the Buddha (Fig. II.5 & II.6). Conch
shells symbols bear religious meanings as well (Fig. II.7). And it seems that it is
via that route that Indian religions, made their way into Southeast Asia. Hence,
when Brahmins arrived, around AD 500, the population already knew of their
beliefs. It is even possible that Brahmins went to Southeast Asia post-conversion
of the inhabitants of the latter.
14. 14
Figure II.5: Rock crystal lion found in Phu Khao Thong. From Chaisuwan, 2011: Figure 4.9
Figure II.6: Carnelian lion found in Ban Don Ta Phet. From Glover & Bellina, 2011: Figure 2.15
Figure II.7: Conch shell seal found in Phu Khao Thong. From Chaisuwan, 2011: Figure 4.11
15. 15
So that the period traditionally associated with Indianization, from AD 500
onwards, that is characterized by the emergence of Indian religions would find its
roots in the previous trading phase (Bronkhorst 2011; De Casparis et al. 1992;
Gupta 2003). Interestingly both Buddhism and Hinduism coexisted everywhere in
Southeast Asia (De Casparis et al. 1992; Stargardt 2003: 111), with a trend for the
rulers to be Hindu which could probably be explained by the magnified status of
the king in Hinduism (Bronkhorst 2011; Coedès 1975; De Casparis et al. 1992;
Guy 2011; Higham, 2001: 9-1) and would certainly have involved complying to
some Indian rules. Moreover, Brahmins were not only seen as priests but also as
ministers, counsellors, administrators… and surely transmitted Indian conceptions
of the state as well (Coedès 1975; De Casparis et al. 1992). This could explain the
importance of Sanskrit as well that became necessary to designate typical
features of the states according to conceptions not existing in local cultures and
languages.
Indianization, if we may call it like that, would have occurred then through the
Vaishya preparing the arrival of the Brahmins through the diffusion of artefacts that
were bearing some cultural and religious ideas. Brahmins would have been able to
apply Indian conceptions of the state in these countries, partly relying on religious
principles.
Conclusion
As we have seen the term Indianization is not appropriate and we should call it
a sphere of common cultural background. Since it implied the sharing of
technologies, it is not unreasonable to think it might have had an influence on
water management systems. However the difference is the transportability of the
technology, such a connection would imply some skilled Indian travelling to
Southeast Asia or the other way round.
16. 16
B. Funan
Funan being one of the major topics of this dissertation it obviously deserved a
place here. If the name in itself has been long known as the one of a major
Southeast Asian polity involved in trade (Christie 1979: 286; Hall 1999: 192-196;
Pelliot 1903), understanding what it designated exactly and where it was is quite
difficult.
In the following part I will not address all the current debates surrounding it but
I will rather focus on presenting the most accepted ideas on the state and some
debates particularly relevant to our topic.
1. A polity known thanks to China
The word Funan in itself is a Chinese name that would be a transcription of the
Old Khmer word vnam or bnam, meaning “mountain” (Coedès 1975: 36; Ferlus
2005: 3-4; Jacques 1979a: 374-375; 2004: 68; Vickery 2003: 125-126). Indeed
Chinese records are crucial for the knowledge of the area to the extent we have to
rely on their spelling for the name of the polity; it is only from AD 611 Old Khmer
inscriptions will inform us in a direct way (Stark 1998: 178 & 188-189; Vickery
2003: 128-129).
The first occurrence of the name comes from the Sanguo Zhi written at the
end of the 3rd century CE. It tells the story of two envoys sent to the polity between
AD 245 and AD 250, Kang Tai and Zhu Ying (Jacques 1986: 63; Pelliot 1903: 251
& 303; Stark 1998: 179-180; Vickery 2003: 105). Their mission happened after
Fan Chan, Funan’s king, sent an embassy to China in AD 243 (Coedès 1975: 40-
41; Jacques 2004: 68); thus the Funanese were the first to make contact (Hall
1999: 186; Jacques 2004: 69). These first Chinese ambassadors made several
descriptions of the polity and gave us the first information we possess: taxes were
paid in perfumes, gold, silver… (Jacques 2004: 76) and they had a writing “like the
one of the Hu”1 (Hall 1999: 194; Higham 2001: 24; 2013: 102; Ishizawa 1995: 17).
1
Which is Indian-like
17. 17
However they were not mere visiting officials: it was at the time of the Three
Kingdoms (AD 220-265), and this embassy was coming from the Wu Empire, the
only one to which access to the terrestrial Silk Road was denied due to its location
(Ishizawa 1995). Hence if they emphasized the trading capacity of Funan it need
not mean that this was the only particularity of the polity. Furthermore Fan Chan
had also sent an embassy to India around AD 225 (Coedès 1975: 46 -47; Jacques
2004: 72-74; Pelliot 1903: 271; Vickery 2003: 108) that had just returned with
some foreign goods and envoys; we can be sure the Chinese and the Indian
envoys met in Funan, perhaps partly explaining the success of the polity in China
later, knowing it had contacts with both ends of the Asian Silk Road (Pelliot 1903:
276-278). From this point numerous embassies from Funan were sent to China,
paying tribute from what is written in the annals. However the notions of tribute
and vassalisation are difficult to appreciate as I will develop in part II.C. These
embassies were above all important to maintain some trading contacts (Jacques
2004: 68).
Those Chinese records are also the only indication we have about the location
of Funan. They said that in Funan was a river which was flowing from the west to
the east, which fits the course of the Mekong or of the Bassac (Pelliot 1903).
Moreover it had a border with Lin-Yi, kingdom located in North and Central
Vietnam (Jacques 2004: 68; Pelliot 1903; Vickery 2003: 121). It is thus widely
accepted by scholars that the political heart of Funan was in the Lower Mekong
Delta (Jacques 2004: 6; Pelliot 1903; Vickery 2003: 121; 8). Archaeological
excavations led in Oc Eo, in actual Vietnam, unearthed features that seem to
match the descriptions given by the Chinese such, as jewellery workshops
(Higham 2001: 26; Jacques 2004; 77-83).
When the state was formed is unknown. However from 111 BC, when China
took over the region of Tonkin and North Vietnam, exchanges between Southeast
Asia and China increased, (Ishizawa 1995; Stark 1998: 180). The name of the
polity was not mentioned around the beginning of the Christian Era in Chinese
records (Ishizawa 1995); hence it is believed it was created somewhere during the
1st century CE (Hall 1999: 192; Higham 2001: 33). Nonetheless from the evidence
we have in the Mekong Delta, the occupation of many settlements began
somewhere in the middle of the first millennium BCE, hinting Funan might have
18. 18
been created before the 1st century CE (Stark 2006), although perhaps not as a
coherent entity.
The location of Funan’s capital is unknown as well. According to scholars the
Chinese recorded the name Tomou, possibly equivalent to Vyâdhapura, and it has
first been suggested this was Ba Phnom (Coedès 1943; 1975: 36-37; Higham
2001: 24-25; Jacques 2004: 83). However the distance of 500 li from the sea was
recorded for the location of the capital, which would match more accurately Angkor
Borei, supposition supported by archaeological excavations (Higham 2001: 24-25
& 28-29; Jacques 1986: 62; 2004: 83-90; Pelliot 1903: 289-290; Stark 1998; 2006:
106; Vickery, 2003: 126-131).
Finally, we might enquire about the boundaries of the polity. Around 225,
Funan seemed to occupy all of the Mekong Delta and the coast to the Kra Isthmus
after the conquest of the general Fan Shih Man (Coedès 1975: 36-38; Ishizawa
1995: 14; Jacques 1986: 62; 1995: 36; 2004: 68-69; Vickery 2003: 136).
Nevertheless it was probably only a succession of Funanese ports on the coast
from the Delta to the Isthmus without necessarily going inland (Allen 1998;
Jacques 1986: 61-62; 2004: 74). Furthermore it seems it contracted shortly after
this phase (Jacques 2004; 74).
I will not go through the whole chronology of Funan that has been developed
elsewhere (Coedès 1975; Higham 2001; Jacques 1986; 1995; Pelliot 1903;);
however it has to be noted a Kaundinya Jayavarman (480 - 514) sent several
embassies to China, among which one to ask for some help against Lin-Yi in 484
(Coedès 1975: 57-59; Pelliot 1903: 257-261 & 294;). China denied any help but
made him “General of the pacified South, king of Funan” in 503 (Coedès 1975: 58-
59; Ishizawa 1995: 17; Pelliot 1903: 269 & 294; Vickery 2003: 115). Interestingly
they gave the same title to the king of Lin-Yi around the same time (Vickery 2003:
115). It seems Funan was losing some of its power already at that time. This is
supported by the last recorded king of Funan in Chinese annals, Rudravarman,
who ruled from 514 to around 540-550, just after Kaundinya Jayavarman (Coedès
1975: 60-61 & 66-68; Jacques 1986: 63-64; 2004: 83; Pelliot 1903: esp. 296 &
298).
However the last embassy to China was sent much later, in 649, while Chenla
was already occupying the main territory (Higham 2001: 25).
19. 19
2. The debates surrounding Funan
An important debate is the status of the kingdom. For many academics the
Chinese tried to apply to an unknown system words they usually used (Stark 1998:
186). Thus it is possible Funan was not a kingdom but rather a patchwork of many
realms that were only occasionally assembled in a coherent entity (Hall 1999: 193-
194; Jacques 1979a: 378; 1986: 90; 1995: 33;). However we do have records of
conquests such as the one of Fan Shih Man that leave some stains in this theory.
Either he was an exception, or his conquest would have mainly involved making
the others his vassals.
Another one is without a doubt that of the ethnicity of the Funanese (Jacques
2004: 68; Vickery 2003: 122-125). From 611 AD Khmer inscriptions inform us on it
but not before. An argument that could lead us to think they were already Khmers
is a remarkable continuity in the record between Funan and the subsequent
periods (Jacques 2004: 65-66 & 77-83; Stark 2006). Some scholars say as well
that if the name Funan was derived from a Khmer word, its inhabitants ought to be
Khmer (Jacques 1979a: 374-375).
Finally and particularly relevant to our topic is the debate surrounding the
organization of Funan. Most archaeologists argue that it was divided in two big
parts, one on the coasts that was concerned with trade and one in the hinterlands
that would produce the necessary agricultural surplus (Hall 1999: 190 & 192-193;
Ishizawa 1995: 15 & 17; Stark 1998: 185; 2006: 99). Hence we would expect to
see very different water management systems between these two areas.
Conclusion
Funan is certainly a difficult-to-reach polity. However given its historical
importance according to the Chinese and being the only hint we have for that time
in the Delta, it is an important one.
20. 20
C. Chenla
Chenla is a name appearing in the Chinese annals at the occasion of the first
embassy sent by this polity in 616 under the king Ishanavarman (Coedès 1943;
1975: 69; Pelliot 1903: from 272; Vickery 2003: 131). According to these, Chenla
was a vassal of Funan that managed to overtake its suzerain (Coedès 1943: 1;
Jacques 1986, 61; 2004: 95). However, fortunately, this is a part of history for
which we have numerous inscriptions that allow us to fill in the gaps left by these
records (Higham 2013: 108; Stark 1998;). These inscriptions, in Sanskrit and
Khmer, also indicate that Chenla had a significant Khmer component.
In the subsequent part I will develop Chenla’s relationships to Funan and
China before addressing the creation of Chenla and its separation into two entities.
1. Chenla, Funan and China
First of all it is important to note some authors have interpreted Chenla has
being another name for Funan. I disagree with this idea since we have the record
of the two polities sending embassies to China in the same years (Pelliot 1903;
Vickery 2003: 131). It is unlikely they would have used different names for the
same country. Moreover, the first time the Chinese mentioned Chenla, in the
“History of the Souei” (581-618), they wrote that the king Citrasena, followed by his
son Ishanavarman, conquered Funan and that the capital of the latter had to move
south to Na-Fu-Na (which has yet to be identified). They add that Chenla was
previously a vassal of Chenla (Coedès 1943; 1975: 65; Ishizawa 1995: 16;
Jacques 1979a; 1986: 61; Pelliot 1903; Vickery 2003: 131-132). The fact that they
cite a new name for the capital of Funan definitely shows that Funan and Chenla
were distinct entities.
It has also been said sometimes that Chenla had never been a vassal of
Funan but that it was a Chinese conception (Jacques 2004: 95-98); however it
could be that this information came from the envoys from Chenla. It has to be
21. 21
reminded here the patron-client system that still exists in Cambodia and that was
certainly important at that time (see Hanks Jr. 1962; Scott 1977; Ayres 2003 for
further discussion on patronage). The client seeks the most important patron in
order to have his protection and gives something in exchange. The patron is
looking for the greatest number of clients and for the most prestigious ones; the
more clients he has and the more prestigious they are, the more powerful he
becomes. Thus the hierarchy among two persons is made according to who is
their patron and who are their clients. The system of vassalisation was probably
similar to a patron-client relationship on the city-state scale. Hence being a vassal
can be positive and it is certainly this kind of factors that encouraged the kings of
Funan and Chenla to “pay tribute” to China. Moreover, being vassal of China
placed Chenla at the same level as Funan; embassies could have been sent to
China to seek legitimacy from its Emperor.
2. Creation of the kingdom of Chenla
Nevertheless, if Citrasena is mentioned as the first king of Chenla in the
Chinese annals, we know this was not the case sensu stricto.
To understand this we need to return to his father we know thanks to
inscriptions, Vîravarman. We notably find his trace in an inscription from Wat Phu
that suggests he was the king of this city in the 6th century (Coedès 1975: 65-66;
Jacques 1986; 2004: 96-97; Santoni & Hawixbrock 1998: 388). Furthermore the
city would have been built in the end of the 5th century, antedating from a century
the apparition of Chenla. Apparently, while Citrasena ruled the city after the death
of his father (Santoni & Hawixbrock 1998: 388), Bhavavarman I, his brother, had
been denied inheritance and created his own empire doing warfare, the heart of
which was located in Sambor Prei Kuk where he was ruling in 598 (Coedès 1943;
1975: 68-69; Jacques 1986; 2004: 118-119). It is sure Bhavavarman I went as far
south as Battambang, which would have under Funan’s influence. In the meantime
Citrasena extended his empire to the west into the Thai Mun valley (Coedès 1975:
68-69; Higham 2001: 38; 2013: 106 & 109; Jacques 1986: 69). When
Bhavavarman I died, Citrasena took his throne and picked the name
Mahendravarman (Jacques 1986; Pelliot 1903: 295-296). Thus he was the first to
22. 22
rule on both kingdoms and it is this reunion that has been called Chenla.
Mahendravarman’s son, Ishanavarman ruled over both kingdoms as well from 616
at the latest and died around 637, when Funan totally disappeared (Coedès 1943:
1; 1975: 69-70; Higham 2001: 38-39; Ishizawa 1995: 16; Jacques 2004: 134;
Pelliot 1903: 248).
It has sometimes been suggested from inscriptions (K. 286 for instance) that
Mahendravarman was Rudravarman’s grandson, but this seems inaccurate and
no blood connection can be made between them (Coedès 1975: 66-67; Jacques
1979a; 1986: 64; 1995: 34; 2004: 97; Vickery 2003: 134-135). However an
inscription from Kdei Ang (K. 53), dated AD 667, tells us about a family of retainers
that served Rudravarman, Bhavavarman I, Mahendravarman, Ishanavarman and
Jayavarman I (Higham 2001: 37 & 41; Vickery 2003: 134-135), hinting the
transition from Funan to Chenla has been smooth for many people despite the
possible conquest of Bhavavarman I, which is consistent with the continuity in the
archaeological record.
After this extension, the kingdom contracted until the reign of Jayavarman I
(635-680) who resumed it to occupy the whole delta (Coedès 1943: 5; Higham
2001: 39-42; 2013: 109-110; Jacques 1986; 2004: 109 & 139). He developed the
administration (Higham 2001: 40-52) and created official charges such as the chief
of the royal grain stores (dhanyakarapati) (Higham 2001: 41; 2013: 109-110) and
changed the charge of pon that was inherited by the mratan he designated himself
(Higham 2001: 44; 2013: 110; Jacques 1986: 89; Vickery 2003: 117); thus
tightening his grip on his kingdom.
3. Water Chenla and Land Chenla
Even though, Chenla divided after Jayavarman’s death in what has been
called Water Chenla and Land Chenla (Jacques 1986: 59; 1995; 2004: 139; Pelliot
1903: 286). We do not know what each of these encompassed and no borders
delimiting them have been traced but the former would probably be inland and the
latter on the coast (Jacques 1986; 1995; 2004: 146). Even the date of the
separation is unknown; Jacques (1986: 85) suggests 707 and Coedès (1975: 85)
23. 23
says after 706. Moreover, from the epigraphy dozens of small realms existed in
the area (Jacques 1979a: 378), so that archaeologists argue there has never been
such separation between two entities (Coedès 1975: 86). However it could have
been a way of separating two different kinds of cultures among those small states;
the Land mountainous Chenla and the maritime coastal Chenla.
In 713 one of these states would have been led by the daughter of
Jayavarman I, Jayadevi, who established her city in the area that would later be
the one of Angkor (Coedès 1975: 85; Higham 2001: 42; Jacques 2004: 139).
Since her daughter reigned, Jayavarman I had probably no heir (Coedès 1975:
72). Maybe in a phallic society dominated by Hinduism the supremacy of a woman
could not have been accepted so that the empire divided?
This period of separation lasted until around 800 CE, when Jayavarman II
started unifying the kingdoms (Coedès 1975: 97-103; Higham 2013: 111-112;
Vickery 2003: 133-134). He declared himself universal ruler in 802 and installed
his capital in Angkor; next to the previous capital of Jayadevi with whom he
claimed some connection.
Conclusion
From the evidence it seems Chenla appeared before the end of Funan,
centred on Wat Phu. Both probably used to be competing polities. However the
name Chenla, that is Chinese (Coedès 1975: 65; Jacques 1979a: 371; 1986: 60;
1995: 33; 2004: 96) would have been given only after the kingdom started to
significantly shrink parts of Funan. It is certainly within this period, for which the
inscriptions are widely dispersed, that we see the roots of the kingdom of Angkor.
One of Chenla’s main features would have been the unification of very different
landscapes, some mountainous and some flat flooded plains, in a single entity.
Perhaps it is a reason of its instability.
24. 24
Chapter conclusion
In virtue of the sphere of shared technical knowledge we have addressed, we
can legitimately ask if there are some connections between the water
managements systems from Funan and Chenla. Nevertheless, Funan would have
been divided into two big parts, one concerned with trade and the other one with
the production of the agrarian surplus, so that we could expect different water
management systems. Chenla, in turn, was more turned towards the hinterland
since they had no access to the sea in the beginning and we do not have much
record of them doing trade so that we would expect to see more complex water
management systems. Overall a remarkable cultural continuity seems to
characterize the archaeological record (Jacques 2004: 65-66; Stark 2006), so that
it is possible that all these wars and changes were mainly a ruler’s concern and
that in the end they did not influence much the commoners and the way things
were done.
25. 25
III.The use of water
The most obvious use of water is for personal consumption such as drinking
and cooking. However as we will see in the following parts, water is important for
many other things, such as agriculture, transportation and religion.
A. Agriculture: rice
The most common kind of crop encountered in the whole of Asia is without
doubt rice (Capistrano & Marten 1986). In the studied area, rice grows naturally in
its wild state (Khush 1997: 27-29; Fuller & Qin 2009: 89-94; Fuller 2011: 79-80)
(Fig. III.1). However it would be far too easy to assume that the domestication was
local and instead it is a far more complicated picture that emerges when we pay
interest to the culture of the cereal.
Figure III.1: Map of the zone occupied by wild rice during the Last Glacial (20,000 BP) and its expansion in the Early
Holocene (9,000 BP). From Fuller 2011: Fig.1
26. 26
The first question is, was there already rice production in Funan? And for once
the answer is quite easy as it is a resounding yes since the envoy of Kang Tai we
have already mentioned wrote rice was cultivated in Funan (Higham 2001: 33).
However he gives us no clue about the way it was cultivated and even less about
the breed that was used.
1. Growing rice
First let us begin with the ways of growing rice for each requires different kinds
of water management systems. Usually it is considered there are two kinds of rice,
wet rice and dry rice; the difference is only the way in which they are grown,
respectively in inundated fields or during the dry season in non-inundated ones
(Capistrano & Marten 1986: 11). Then different irrigation or non-irrigation
techniques can intervene. Finally it has to be noted that it is usual to practice
swidden agriculture in Southeast Asia (Capistrano & Marten 1986: 13-14;
Srimongkol & Marten 1986). This technique consists in burning a patch of forest to
sow it, the ashes acting as natural fertilizer. After a couple of years the area is left
fallow for sometimes as much as a decade and another patch nearby is burnt. This
technique is usually associated with dry rice agriculture but can be adapted to
others types of cultivation (Capistrano & Marten 1986: 13-14; Srimongkol & Marten
1986).
The most common way of growing rice is without a doubt in inundated paddies
(see Capistrano & Marten 1986; Khush 1997: 31; Lando & Mak 1994; Srimongkol
& Marten 1986; Van Lière 1980). In that case rice is first put to germinate in
nurseries before being transplanted in inundated bunded-fields during the wet
season, often around July to avoid any drought. Generally this is made in low-lying
lands. This system is often rainfed, meaning that the precipitations of the monsoon
are sufficient to ensure the level of water in the paddy. However water cannot be
too high, usually about 50 cm is required; hence some devices have to be used to
drain the excess of water. It can be a terrace system for instance where the
27. 27
excess of water goes above the border of the field (Stargardt 1977). On the other
hand it does not like having much less; that is when irrigation techniques can
occur. Moreover irrigated rice can be grown all year-round (Capistrano & Marten
1986: 11). Finally in this case rice has higher yields if the surface of the water is
perturbed; some techniques have been developed to ensure it (Fuller & Qin 2009).
In the case of lowlands that are seasonally inundated, such as the Mekong Delta,
the level of water is naturally sufficient and only two conditions are important,
knowing the pace of the inundation to sow and transplant at the right time and
knowing the maximum depth of water, otherwise the culture is lost; moreover the
seeds can be broadcast and not transplanted, requiring less labour and no special
water management nor terracing (Van Lière 1980: 267-271). In Southeast Asia it is
often assumed inundated rice preceded bunded-fields that most likely require
water management (Van Lière 1980: 267-271). This technique has the higher
yields.
Another technique that is commonly seen is the upland dry rice (see
Capistrano & Marten 1986; Khush 1997: 32; Polthanee & Marten 1986). Rice is
sown towards the end of the wet season and grows thanks to the natural residual
moisture. Of course this requires a soil having sufficient water holding capacity
and the yields are inferior to the one of the paddies (Khush 1997: 31-32). In some
cases water can be lifted to ensure sufficient moisture. Not much water is required
but the yields are low.
Deepwater rice is also recurrent (see Khush 1997: 32; Lando & Mak 1994).
Chou Takuan, in the 13th century, wrote that this kind of rice grew naturally around
the Tonle Sap (Lando & Mak 1994: 3). Usually, to be considered as deepwater,
rice has to grow in more than a meter of water. However the key lies more in its
elongation capacity. The straws of the previous year are burnt in the field and the
rice is sown in dry earth before the beginning of the inundation. It starts
germinating. When the time of the flooding comes, elongation of the straw is
stimulated by the rise of water in order to keep leaves above water. Interestingly
some species falling in the deepwater type for they grow in more than 1 meter of
water and have elongation capacity are categorized as rainfed lowland by
Cambodian farmers (Lando & Mak 1994: 12-14). For them deepwater rice is only
to be sold and not for personal consumption (Lando & Mak 1994: 24-26),
28. 28
indicating a low cultural value of this rice. Rainfed lowland rice possessing some
elongation properties could be some breeds that have been hybridized with
deepwater rice. The biggest advantage of deepwater cultivation is that no water
management is required, the cons are that a full field can be easily lost in case of
drought or too important inundations; hence the environment has to be precisely
known. The yields are below the ones of lowland rainfed rice but above the dry
rice ones (Lando & Mak 1994).
Finally, the last important type of culture of rice is the flood-recession kind (see
Fox & Ledgerwood 1999; Van Lière 1980). This system consists in planting dry
rice after the inundation. It is often transplanted as soon as the floods retreat to
benefit from the deposits they leave. It also has to be irrigated 2 to 3 weeks after
having been transplanted so that it does not dry out. For this purpose some
temporary dykes are built upland to store some water during the recession of the
flood. With this method it appears ploughing is not necessary for the soil is rich
enough in nutrients thanks to inundation. The yields are similar to the ones of
irrigated wet rice (Fox & Ledgerwood 1999). Some archaeologists believe the
practice of flood-recession agriculture could be a viable explanation for the system
of canals seen in the Delta (Fox & Ledgerwood 1999; Higham 2001: 33). However
its use is not at all confirmed in previous phases (Fox & Ledgerwood 1999; Van
Lière 1980).
2. A short history of the domestication of rice and its dispersal
Indeed this is a presentation of the traditional ways of cultivating rice from a
modern point of view. But their ancestry at the time of Funan and Chenla are less
than known. If rice was already grown in paddies in China, it was probably not the
case in Southeast Asia. To understand the way cultivation could have been
practiced we need to understand the domestication of rice and its dispersal.
I will only present here the most widely accepted theory for I wish not to enter
in the debate of the potential multi-domestication of rice (Fuller & Qin 2009; Fuller
2011; Van Driem 2012). Two main breeds are commonly cultivated, japonica and
indica (both being variations of oryza sativa, an annual kind of rice that would itself
29. 29
come from the perennial oryza rufipogon for japonica and the annual oryza
rufipogon nivara crossed with japonica for indica) (Bellwood 2011; Fuller & Qin
2009; Fuller 2011; Khush 1997: 27; Van Driem 2012: 118). Nowadays the
cultivated cereal in the area that interests us is indica (Castillo 2011: 116),
however the first that was domesticated was japonica in the Middle Yangzi, in
water managed bunded-fields (Fuller & Qin 2009; Fuller 2011: 83-84; Van Driem
2012: 120-121). Its domestication would have been complete by 4,000 BC
(Bellwood 2011: 95; Fuller & Qin 2009; Fuller 2011: 83-84). From this cradle
japonica would have spread:
- East to the area of Taiwan from where it was subsequently diffused into
insular Southeast Asia around the change of era (Fuller 2011: 85-86; Pejros
& Shnirelman 1998;).
- West to Northwest India. There it would have hybridized, perhaps naturally,
with oryza rufipogon nivara that grows naturally in the area to give birth to
the nowadays indica (Fuller 2011). It seems it would have been completed
around 2,000 BC (Fuller 2011: 80-82, 86-88).
- South to Southeast Asia, through the mountains of Northeast Thailand and
along the coast (Bellwood 2011; Fuller 2011; Pejros & Shnirelman 1998).
As we can see, it was then japonica that was to be found in Southeast Asia, as
confirmed by the DNA analysis of rice found in Vietnam (Bellwood 2011: 98) and
the findings in Thailand (Castillo 2011). This introduction has not been precisely
dated and some have suggested as soon as 2,000 BC (Fuller 2011), but on the
basis of analyses (King et al. 2014) its cultivation as an important subsistence crop
is only ascertained around 1,000 BC in Northeast Thailand (Castillo 2011: 115;
Van Driem 2012: 121). On the other hand rice coming from Vietnam is poorly
known2 and it is sure that around 2,000 BC japonica was present (Bellwood 2011:
97; Van Driem 2012: 121), which would antedate the North Thailand introduction;
hence hinting a previous coastal dispersal (Zhang & Hung 2010). Moreover the
coastal site of Khok Phanom Di in Thailand delivered domesticated rice remains
dated between 2,000 BC and 1,500 BC (Castillo 2011: 115), which is consistent
2 At least to non-Vietnamese speakers as much of their findings are not translated in English. Access to them would
probably allow one to clarify the situation about rice in Vietnam which could be crucial to understand its dispersal in
Southeast Asia.
30. 30
with this kind of dispersal. However the domestication state in China would have
been inundated wet rice culture whereas from the evidence it seems that in
Thailand it spread as upland dry rice (Castillo 2011; Khush 1997: 30). In turn, it is
possible that rice in Khok Phanom Di depended on swamps under an inundated
form (Thompson 1996) and similarly in the Mekong Delta thanks to the annual
flood (Van Lière 1980). The coastal dispersal would then have been facilitated by
the more suitable environment recreating conditions more similar to its
domesticated state from China. Moreover from δ13C analyses (King et al. 2014)
(Fig. III.2, III.3 & III.4), we clearly see that Ban Chiang in the Sakon Nakhon Basin
hosted cultivators relying on rice between 1050 BC and 900 BC but not after; while
in turn the Upper Mun River Valley hosted rice cultivators from 1,000 BC to 900
BC and maybe even 800 BC. It is quite possible they turned to the more suitable
foxtail millet because of low yields or lack of social complexity to organise the
intensive labour required by rice (Castillo 2011; Fuller 2011: 87).
Figure III.2: Table of the burial phases and samples used for the δ13C analyses. BNW = Ban Non Wat; BLK = Ban Lum
Khao; NUL = Noen U-Loke; BC = Ban Chiang. (King et al., 2014: 113: Table 1)
31. 31
Figure III.3: Results from the analyses in the Upper Mun River Valley. The low value represent an alimentation mainly
relying on rice (such as in BA2). We clearly see that rice was present from BA2 to probably BA3b before being
abandoned. It comes back in IA1 in Noen U-Loke but interestingly not in Ban Non Wat. This suggest the area was not
as uniform as presented in many articles. (King et al., 2014: 118: Figure 2)
Figure III.4: results from the analyses in Ban Chiang. Here rice is cultivated in BA1 but never comes back as a major
crop. (King et al., 2014: 118: Figure 3)
32. 32
But then, when did indica spread from India? Noen U-Loke shows rice grew
anew in Thailand from 400 BC (King et al. 2014). Phum Snay (350 BC – 200 AD)
and Phum Sophy (AD 100 – AD 600) seem to confirm rice was grown around that
time in Northwest Cambodia (Newton et al. 2013). It is very tempting to associate
this renewal with an indica arrival; however Ban Non Wat, Phukhao Thong and
Khao Sam Kaeo show that apparently it was still japonica around 400 – 200 BC
and that it was still cultivated in dryland conditions in Thailand (Castillo 2011). It
would then be only a second dispersal. The development of complex moated sites
such as Angkor Borei, along with more complex mortuary patterns in the area, are
definitely contemporary with it (Castillo 2011; Fuller & Qin 2009; Fuller 2011: 87;
Stark 1998). Thus this new dispersal and those new signs are probably the sign
labour could be organised. In fact, indica would have been introduced much later,
around AD 300 for some scholars (Castillo 2011).
Conclusion
To summarize, rice has been domesticated in China in paddy fields and was
introduced in Southeast Asia along the borders of Vietnam around 2,000 BC. We
do not have much information about its dispersal; however rice could have been
cultivated in inundated areas or swamps during the wet season or as following the
retreat of the flood in coastal areas while in the mountains around Thailand it
would have been grown in dry conditions. The production in Northeast Thailand
did not last, perhaps because upland rainfed dry rice did not yield enough but most
likely because of a lack of social complexity allowing a true organization of labour,
on the other hand the methods used along the coast would require far less labour.
Then at the beginning of the Iron Age (ca. 400 BC), there would have been a new
dispersal of japonica rice. Indica would have arrived much later and so would have
irrigated wet rice that are both now the most popular forms. However it is quite
evocative that we have no trace of its arrival, we could have thought it would have
been perceived as a major change by local populations.
33. 33
B. Water transportation
It has been postulated that in the beginning the Malays were the ones having
the boats (Higham 2001; 23-24; Jacques 2004: 65; Vickery 2003: 111-112).
Moreover the dispersal of the Austronesians from the island of Taiwan certainly
involved boats to be able to reach the Philippines for instance. In Funan the
presence of boats is ensured by the testimony of Kang Tai but also by the mission
sent to India by Fan Chan. Kang Tai speaks of boats 46 to 50 meters long, 4.60 to
5 meters high, able to transport as many as 600 to 700 men and 1,000 metric tons
of merchandise that would have got up to 4 sails (Porée-Maspero 1986: 69).
However, all of these were maritime boats and gives us no indication on the
presence of river boats in the area.
Quite recently in Northern Vietnam some boats have been discovered made
with locked mortise-and-tenon technology and dated around the change of era
(Bellwood & Cameron 2007). But more importantly, the Viet Khe boat shows us a
technology in wish the canoe was dugout in a log (Ibid.). This simple dugout would
have been suitable for navigation in canals.
The stitched plank and lashed-lug technique has been extensively studied by
Manguin (1996) and proves typical of Southeast Asia. The first known example of
this technique is the boat recovered in Pontian, Malaysia, and dated around AD
260 – AD 430. One of the important features of these boats would have been to be
very stable on the sea but also to have been able to navigate in canals.
Finally ethnography and literature also confirm canals were been used for
navigation since early time, although not giving us much information about their
specificities (Bellwood & Cameron 2007; Porée-Maspero 1986).
34. 34
C. Water, religion and agriculture
In Hinduism the temple is first seen as a transfiguration of Mount Meru
surrounded by the primordial sea that justified the creation of moats such as in
Angkor Wat or of rectangular reservoirs, the barays, in the Khmer civilization
(Hardy 2007). Water was also particularly important in Shaivism for Shiva is the
God of both destruction and fertility, the latter being closely related to water.
Buddhism also values the presence of water next to the settlements. The
symbolic of purity and clarity is also present but more importantly water is used as
a meditation support.
However, more than the symbolic of water in the religions per se, it is the
concept of monastic landlordism that is particularly relevant to our topic. From a
number of texts and inscriptions we know local rulers and kings donated lands,
villages and workers to Buddhist monasteries, as soon as the 2nd century BC in Sri
Lanka and probably as early in Central India as well (Coningham et al. 2007;
Gunawardana 1979; Morrison 1995; Shaw & Sutcliffe 2003;). These donations
were used to produce merit, the core concept of Buddhism since merit-making is
the only thing the laity can do to improve their karma and hope to leave the
samsara in future lives. However, before the time of these donations and
monasteries, monks were living far from the laity according to the ideal of the
monk in meditation, away from the mundane world, and begged for their food. Two
explanations are traditionally summoned to explain their new sedentary life; either
the monks settled down during the Asokan period due to royal patronage that
established monasteries to create a network used to hold the population together;
or the local population depending more on Buddhism required them to settle down
to be closer to the laity and for religious purposes (the longer a donation is used
the more merit is produced, permanent settlements would then produce merits for
possibly centuries). Having settled down and under the pressure of the growing
monastic community, generating a food surplus would have certainly been needed
(Gunawardana 1979; Shaw & Sutcliffe 2003); furthermore this food surplus also
served the population and would alleviate in some way part of the pain
experienced by the villagers, a way to justify their practice from a Buddhist point of
35. 35
view. Irrigation certainly allows an increase of the yields of the crops (Shaw &
Sutcliffe 2003) and numerous inscriptions celebrating the construction and
maintenance of reservoirs and dams as donations to monasteries seem to confirm
it was used (Gunawardana 1979; Shaw & Sutcliffe 2003). Furthermore, the spatial
relationship between religious establishments and water management structures is
striking and both seem to develop synchronically, thus implying most of the water
management installations were created because of the establishment of the
monasteries (Fogelin 2006; Shaw & Sutcliffe 2003; 2003b). Moreover in places
such as Sanchi traditional wheat crops did not need irrigation thanks to the nature
of the soil; therefore irrigation may be an indicator of a switch to rice agriculture.
Rice would have been brought by the Buddhist monks because of the importance
of the cereal in the religion (Shaw & Sutcliffe 2003). But if irrigation and rice culture
can be linked to the Buddhists monasteries, monks are not allowed to produce
their food, which justified the donations of workers. However since those people
also needed some food and made most of the work, a kind of tax was raised: the
monasteries had a percentage of the crops (Gunawardana 1979).
To summarize, giving food would not have been sufficient to support the
monasteries; hence justifying giving them crops; moreover crops have a longer
uselife thus generating more merit; since high yields were needed irrigation was
used; furthermore irrigation allowed the monks to produce rice and to alleviate the
pain of the laity; monks could not work so that workers were given; monasteries
received a share of the harvest. Hence the monasteries in India were acting as
landlords and dispersal of rice and irrigation were tightly linked to them.
However, this system was not confined to Buddhist institutions and by the 4th
century CE Hinduism had adopted the same landlordism techniques (Shaw et al.
2007: 171-172). That is particularly relevant to our topic since the Chenla period is
the first time we have inscriptions mentioning donations of slaves and rice fields to
Hindu temples (Jacques 1979b); thus it possible to suggest a similar kind of
monastic landlordism would have been present in Cambodia and its surroundings
from at least the 7th century CE.
36. 36
Chapter conclusion
As we have seen, water can be used in a number of different ways that imply
different water management systems: reservoirs for rice irrigation, canals for
transportation… hence analysing the ones present in Funan and Chenla could
help us determine their use.
37. 37
IV. Case study
I will be reviewing here some sites from Funan, Chenla and India in order to
assess their water management systems and try to draw some conclusions on
their evolution and potential mutual influence.
A. Methodology
In the following parts I have chosen two sites for Funan, Angkor Borei and
Noen U-Loke. Angkor Borei certainly is an easy choice since as we have said
previously it probably used to be one of the capitals of Funan. Noen U-Loke is
located in Northeast Thailand and is not per se from Funan, however much of its
occupation can be dated from the same period and it could perhaps be
representative of the agricultural hinterland that was supposed to have supported
the coastal sites of Funan according to the models exposed previously; indeed the
sites from the Upper Mun Valley River among which is Noen U-Loke are similar to
the ones from Northwest Cambodia (Moore et al. 2007).
The two sites for Chenla were even more logical since we have two cities that
have been identified with capitals of this kingdom, Wat Phu in Southern Laos and
Sambor Prei Kuk in Cambodia.
India will be less studied in this dissertation unfortunately, partly because of
the lack of room. Numerous different water management systems existed in India
depending on the natural environment; thus I will be only presenting some of them.
Since no technical study of the water management systems has ever been
made in Cambodia, Laos, or Northeast Thailand, I will review the following in my
case studies:
- Recognizing and describing the water management features. Quite
surprisingly a great number of them have never been studied although well-
known for a long time.
38. 38
- Assessing the way in which they are connected to the network and related
to each other
- Offering a possible explanation for their use
Finally in the next chapter we will be assessing what we can learn about the
developments of these polities thanks to the study of water management sysems
B. Water management systems in Funan
1. Angkor Borei
Angkor Borei is an archaeological site located in the Lower Mekong Delta on
the Cambodian side in the province of Takeo (Stark 1998: 188). Angkor Borei’s
location and possible importance has first been acknowledged in the beginning of
the 20th century (Stark 2006). However it has not been the main focus of
excavations, the Vietnamese side and particularly Oc Eo having been previously
considered more important3. The Khmers Rouges obviously prevented any
archaeological field. Finally it is only in the second half of the 1990s that
excavation started at the site in the frame of the “Lower Mekong Archaeological
Project” (LOMAP) (Stark et al. 1999: 11; Stark & Sovath 2001: 88-94; Stark 2006).
Nowadays a great part of the remains are located under present housing or
paddies, considerably complicating the research (Stark et al. 1999: 10-11).
3
I have voluntarily chosen not to study Oc Eo for reasons I will explain later
39. 39
Figure IV.1: Archaeological map of Angkor Borei after the 1995-1996 excavation. From Stark et al. 1999: Fig. 4
Angkor Borei (Fig. IV.2) is thought to have been first settled during the 4th
century BC (Stark 1998: 193-194; 2006; Stark et al. 1999). Then it would have
known a peak of activity in the 6th century before being less inhabited (Bishop et
al. 2003). However remains from a temple show that it was still inhabited in the
10th century (Stark et al. 1999: 26-27), hinting a continual occupation. The site in
itself is on an elevated platform on the edge of the seasonally inundated area
(Stark 1998: 188). From July to November, it is surrounded by water, except on
the Northwest side (Stark 1998: 188). It is characterized as a walled and moated
40. 40
site of 300 ha (Stark 1998: 191). It possesses a double moat in some places along
with a wall, either between the moats or on the inside when only one moat
appears. The locality is also famous for its connection to the ancient network of
canals mapped by Paris (1931; 1941) and Malleret (1959).
The first part of interest in this study is the 22 meters wide moats. In fact they
are partly natural, created on the North side by a natural river and maybe on the
West side by an ancient channel according (Stark et al. 1999: 20-22; Stark &
Sovath 2001: 89). On the other sides they are far more regular and there are few
doubts if any about their anthropogenic nature. However even these
anthropogenic parts of the moats were clearly connected to the river (Stark 1998:
190-191), suggesting there were filled by the latter, ensuring year-round supply of
water. Finally it has to be noticed that the Angkor Borei river is entering the city on
the North side and exiting on the West side; however there is an interesting
feature connected to the river at the centre of the city that is often inundated and
might have been part of the natural channels. As we can see these moats,
although partly anthropogenic are heavily relying on the natural watercourses and
are above all a matter of channelling. They have not been dated or at least the
samples have been dismissed. In fact, there has been an attempt during the 1995
season but the results were considered as inconsistent (Stark et al. 1999: 12-15).
The two samples are coming from the interface between the sterile soil and the fill
in the moats. The first one, sample 89298, returned a calibrated date at 1 δ of AD
5 – AD 100 while the second one, sample 89299, returned two calibrated dates at
1 δ, 3505 BC – 3425 BC and 3390 BC – 3350 BC. It is possible that there is no
problem with this dating, the first sample would come from the time when the
moats have been excavated and the second one from the sterile soil below, there
might have been a mistake in the sampling. This would be consistent with the idea
that the settlement was first inhabited around the 4th century BCE and that would
also be consistent with other moated sites such as Noen U-Loke.
The use of the moats is unknown. There would have been no point in creating
them for storing water given the environment for there is water all year-round
thanks to the river. Supplying water to the community is also unlikely. Protection is
of course possible as suggested by the walls that seem to be 2.40 meters wide
and 4.50 meters high (Stark et al. 1999: 22-23); at least it could have been a
41. 41
protection against wild animals. Transportation would have been possible although
there would not have been much point in building such complex moats for this
purpose given that the river is going through the settlement. Finally the use for
agriculture is unlikely since there is no redistribution system so that it would have
been extremely localized and would have required other reservoirs in the
lanscape. Hence protection seems to prevail. However religious conceptions might
have been an influence, but this is only speculative.
The second feature of interest in the area is the ponds. They have not been
properly studied nor dated. Nevertheless inside the walls we notice a great
number of them. They are always irregular and not connected to any supply; thus
they were probably rainfed. They would have been most probably used for
personal consumption of water next to the housing. Although there has been no
precise dating of them and the site has been inhabited at least until the 10th
century CE, since the peak of occupation is supposed to have occurred in the 6th
century CE and these ponds are scattered all around the site we can hypothesize
without much risk some of them are from before the 6th century. Apart from those
ponds we see two barays just next to the moats. The barays are usually
associated with later periods and their presence is unusual. The most eastern one
has been dated and it has been suggested it would have been built much later,
around the 17th century CE, on a previous body of standing water (Bishop et al.
2003). That would also be consistent with the idea that the frequently inundated
area in the city we have addressed before used to be a waterway. The other one
has not been dated but I suspect it is contemporary with the eastern one. Hence I
will not discuss further their use. According to Stark (1998: 191; 2006; Stark &
Sovath 2001: 90) there are also some irregular rainfed ponds scattered in the
area, usually associated with mounds, but I could not find any map mentioning
them.
Finally the last striking feature and the one that has first attracted the
archaeologists in the area are the canals (Fig. IV.3). They are everywhere in the
delta, from Angkor Borei to the sea. Around a couple of settlements it seems there
is a radiating system of canals: it is the case of Angkor Borei or Oc Eo (Paris 1931;
1941; Stark & Sovath 2001: 91). Their chronology is not well established,
nonetheless two of them have been dated around Angkor Borei: the canal 4 that is
42. 42
around 80 km long and connects Angkor Borei to Oc Eo is estimated to have been
in use between the mid-first millennium BC and the 8th or 9th century CE
(Sanderson et al. 2007); the canal 2 is far shorter and links Angkor Borei with
Kampong Youl and was in use between the mid-first millennium BC and the early
5th to early 6th century (Bishop et al. 2004), hence when the site is starting to
decline. These data suggest an early excavation for the canals compared to what
we could expect if Funan really started in the 1st century CE, moreover they have
been used mainly at the same time. Canal 4 was about 1.30 meters deep and 70
meters wide were the trench has been made (Sanderson et al. 2007) and canal 2
was 1.20 meters deep and 15.75 meters wide at the location of the trench (Bishop
et al. 2004). Interestingly recent dating (Manguin & Vo Si Khai 2000) has shown
the main development of the canals in Oc Eo started around the late 5th- early 6th
century CE, hence when Angkor Borei would have been already on the edge of its
decline. Perhaps this could be linked with the emergence of Chenla that led the
people from Funan to go South, such as in Oc Eo. They would have then
developed the systems they used in Angkor Borei. It is very important as well to
underline Bourdonneau’s work (2003) and others’ (Bishop et al. 2004; Sanderson
et al. 2007) that show these canals were tightly linked to palaeochannels (Fig.
IV.4). Hence these canals have been incorporated to the natural network right
away. The aim was to bring the water were it was needed by deviating it thanks to
canals.
However the use of the canals once again is unknown. The most probable is
that they would have been useful for transportation since they seem to link
archaeological sites, which seems confirmed by the finding of a paddle in the
canal going through Oc Eo and dated AD 120-390 (Bourdonneau 2003: 270) and
by the size of the canals. The important width of canal 4 while its depth is similar to
the one of canal 2 could then suggest it was a main axis and that numerous boats
needed to be able to navigate at the same time. Another possible use would have
been for agriculture. Nevertheless irrigated agriculture is unlikely for the area is
already inundated part of the year and water is not missing. Furthermore no
evidence point toward irrigated agriculture. Another possibility would be that the
canals were also used to drain away the water when the flood retreats
43. 43
(Bourdonneau 2003; Mabbett & Chandler 1995: 68) which would be consistent
with flood-recession cultivation.
According to results coming from the coring of the eastern baray (Bishop et al.
2003) it seems the environment was of grassland during Funan, perhaps burnt
regularly, while from the 6th century the evidence disappear and there is a
regrowth of the forest. This suggests two things: the first one that agriculture was
practiced around Angkor Borei and that the Delta might not have been as
dependent of the hinterland for its production as sometimes hinted; the second
one that something happened somewhere in the 6th century that led them to stop
their agricultural practices. Nevertheless the site was not abandoned and
agriculture was still probably practiced to sustain them, although perhaps
differently.
44. 44
Figure IV.2: Canals recognized after aerial observations in the middle of the 20th century. In blue actual canals, in red
solid lines ancient canals, in red dashed lines possible ancient canals. From Malleret 1959: Plate 12
45. 45
Figure IV.3: map of the canals around Angkor Borei showing the palaeochannels. From Bishop et al. 2004: Fig. 3
Angkor Borei possesses a water management system heavily depending on
the natural network: the moats are partly created thanks to the river and the canals
are linking paleochannels. Interestingly the canals could predate the moats. The
storage of water, mainly for personal consumption, seems to be consisting in the
collection of rain water. As for agriculture, the record shows its practice changed in
the 6th century CE..
46. 46
2. Noen U-Loke
The second site we are going to study is Noen U-Loke, in the Upper Mun
River Valley, on the Khorat Plateau, in Thailand (Boyd & Habberfield-Short 2007).
I have chosen Noen U-Loke among others from the Upper Mun River Valley
because it is probably one of the better studied and one of the best chronologically
phased.
Noen U-Loke is a 450 meters (E-W) by 250 meters mound that is 5 meters
high in a floodplain in which flows the Mun River (Boyd & Habberfield-Short 2007;
Habberfield-Short & Boyd 2007: 55). In the Iron Age (500 BC – AD 500) it seems
the area hosted many sites that probably had contacts between them (Boyd &
Habberfield-Short 2007: 2). It also seems the climate has changed a lot in this
zone, getting always dryer (Boyd & Habberfield-Short 2007), which has changed
significantly the courses of the river and of its tributaries (Boyd et al. 1999a).
In fact, we have very few water management systems at these sites of
Northeast Thailand. In Noen U-Loke for instance the only ascertained feature is a
succession of 5 moats (Fig. Iv.5 & IV.6). These moats are quite simple but their
construction as inferred from archaeological evidence is quite remarkable and
needs to be developed here. As I have said it appears the tributaries of the river
changed a lot. Apparently when men first settled in the valley, it was much wetter
with “deep single-channel rivers” spanning from 3,500 BP to 2,500 BP (Boyd &
Habberfield-Short 2007: 25-26). In the following period (2,500 BP – 2,200 BP)
many small watercourses were meandering in the valley for the weather was
getting drier (Boyd & Habberfield-Short 2007). Several of these seemed to go
around the mound on which we encounter the site of Noen U-Loke (Boyd et al.
1999a). That would mean that when men first established in the area the mound
on which they set foot was naturally surrounded by water. This is very important in
the choice as it reminds Angkor Borei of course that used the natural setting to
create its moats. Then as we go in time the small channels would have dried
totally and be replaced with anthropogenic moats (Boyd & Habberfield-Short
2007). It had been postulated before that the moats where entirely natural and
solely due to the topography (Boyd et al. 1999b), however when these channels
47. 47
dried Noen U-Loke probably lost its surrounding water. It is then that the mound
would have extended, getting over a few previous channels, and that they would
have created the proper regular moats that we see now. These moats are usually
based on several previous channels being filled, usually covering 2 or 3 previous
channel beds (Boyd et al. 1999b). They have been excavated and dated, hinting
they have been in use probably from the 1st CE to the 6th CE (McGrath & Boyd
2001), hence possibly at the same time as Angkor Borei’s. How these moats were
created is highly speculative, however it has been proposed men created small
channels and let the flow during the wet season erode and enlarge the moats;
hence requiring few people and not much time to dig them (McGrath & Boyd
2001). Unfortunately we do not have any evidence for these practices. However
the fact they followed previous channels might be significant; I would suggest they
did it because following those would ensure the flow to go around the mound.
Following previous patterns would be more likely to recreate previous conditions.
Figure IV.4: Aerial photography of Noen U-Loke. The site in itself is at the centre of the picture, among the concentric
formations. There seem to be a canal but that has not been dated. From Higham & Thosarat, 2007: Fig. 4.3
48. 48
Figure IV.5: Drawing of the mound of Noen U-Loke and of its moats. From Higham & Thosarat 2007: Fig. 4.1
What these moats were used for is also speculative. Defensive moats would of
course come to mind; possibly supported by the density of sites in the area that
could imply some competition for agricultural lands. It has also been proposed it
would have been used as a water reserve, for personal consumption of course,
but also for potential irrigation (Higham et al. 2007). However, although Higham
(2001) has suggested irrigation might have occurred in these sites with have no
evidence such as canal networks. We are sure that the people of Noen U-Loke
had rice, since numerous ricebeds have been found in graves (Talbot 2007), and
that they most probably cultivated it. However Castillo (2011) has shown that rice
agriculture was more likely to have been rainfed dry upland rice thanks to
archaeobotanical data. Moreover the analysis of the evolution of the general
environment (Boyd & McGrath 2001) clearly suggests swidden agriculture usually
more associated with dry cultivation. One of the possible defensive sides might
have been quite simply for the wild animals. We have evidence of tigers in the
area thanks to tiger teeth found in mortuary context (Talbot 2007); hence a moated
49. 49
site could be a solution since tigers usually try to avoid swimming. Transportation
might also have been a possibility as in a rugged area water transportation is
usually the fastest and safest way of travelling. Moreover it looks like the sites of
the area were closely related to one another.
In Noen U-Loke we facing a mounded site, occupied during a time “moats”
were natural and where the inhabitants probably created them anew when they
disappeared. Their use for agricultural practices is highly hypothetical and highly
unlikely given the archaeobotanical data and protection and transportation might
have been possible. But what is more interesting is that we find a situation that is
similar to the one in Angkor Borei: people used minor tributaries to channel the
water where they needed it, probably using canals as well in Noen U-Loke, at least
in the beginning. The creation of canals between the sites would not have been
necessary since the river could already act as such; nevertheless there could have
been a canal visible on the photographs but not dated to my knowledge. Of course
the extent of the system is not as important as the one in Angkor Borei but the
principle is similar.
Conclusion
If we summarize, both in Noen U-Loke and Angkor Borei the systems are
heavily depending on the natural pre-existing network. What has been done could
probably be more likely assimilated to integration into the network and not
management per se. The only storing facilities seem to be for personal
consumption as small and irregular rainfed ponds.
50. 50
C. Water management systems in Chenla
1. Wat Phu
Wat Phu is located in Southern Laos, next to the borders of both Thailand and
Cambodia (Santoni 1997: 1). Although the site has been known from a long time it
has been excavated only from the 1990s by the P.R.A.L. (Projet de Recherches
en Archéologie Lao) before being declared World Heritage site (Santoni et al.
1997: 234; Santoni & Hawixbrok 1998: 387; Santoni 2008).
Wat Phu is a site divided into two sections, a sanctuary and a city (Santoni &
Hawixbrok 1998: 387). To the West there is a famous mountain on which seats a
natural linga that gave the former the name lingaparvata (Defente et al. 1991: 194;
Santoni 1997: 1; 2008: 81; Santoni et al. 1997; Santoni & Hawixbrok 1998: 387). A
natural source is at the base of the mountain, the water of which is considered
sacred (Finot 1902; Santoni & Hawixbrok 1998: 387-388). The sanctuary is at the
base of this mountain as well. Nevertheless the actual sanctuary along with the
two big barays has been created in the 11th century CE, in the Angkorian period
(Defente et al. 1991: 195; Finot 1902; Santoni 1997: 2; 2008; Santoni et al. 1997:
233). Furthermore an ancient road links the sanctuary and the city of Angkor
(Santoni & Hawixbrok 1998: 387; Santoni 2008: 81). There is strong evidence for a
previous pre-Angkorian sanctuary beneath it and it is quite probable the city has
been built here because of the sacred mountain as it is supposed to have been the
case if we consider the version in which a King Devanika established it in the late
5th century (Coedès 1975: 65; Defente et al. 1991: 195; Jacques 1986; 2004: 96-
97; Santoni 1997: 2; Santoni et al. 1997: 234-242; Santoni & Hawixbrok 1998:
387-388). However with no further knowledge about its possible structure it should
be excluded from this part.
The city in itself is located 5 km East of the sanctuary, next to the Mekong
(Santoni & Hawixbrok 1998: 387). Between the sanctuary and the city we find a
floodplain that shows a very dense occupation and notably traces of irrigation
(Defente et al. 1991: 194; Santoni 1997: 1; Santoni & Hawixbrok 1998: 387;
51. 51
Santoni et al. 1997: 233). The city seems to be inhabited between the mid-5th
century and the 7th century, hence in the Chenla period and just before (Santoni
1997: 20; 2008; Santoni et al. 1997). There is a double earthen wall surrounding
the city on the West, South, and probably North sides (The North side would have
been destroyed by the river Houay Sahoua) that would date from the 6th or 7th
century as well (Santoni et al. 1997: 243; Santoni & Hawixbrok 1998: 387-388;
Santoni 2008), to the East we find the Mekong. The walls have been the first ones
to attract the attention after aerial surveys since these features are striking (Fig.
IV.7), moreover they nearly form a square with a 2 km side (2.3 x 1.8 km) (Defente
et al. 1991: 194; Santoni 1997: 5; 2008: 90; Santoni & Hawixbrok 1998: 387-388;
Souksavatdy 1997: 21;). In fact it is a city that is geometric and not depending on
the natural environment, opposite to the ones from the Funan period we have
seen before. It is very possible some religious conceptions dictated this display.
The space between the walls is called the moats but there is a lack of evidence to
be sure they were moats (Santoni & Hawixbrok 1998: 387-388; Santoni 2008). In
fact the presence of features in them that are contemporary with the use of the city
makes it unlikely. Several walls are found inside the city for reasons that are not
yet clear (Santoni 1997). Of particular interest is the small one to the East that
appears to be from the 10th century and probably occupied by Chams (Santoni et
al. 1997: 244). However it is the only element from the 10th century in the city; it
seems that the whole city has been abandoned after the Chenla period and re-
occupied in the 10th century by the Chams in a small part of it. The East side is still
problematic for another reason, the presence of the Mekong. The Mekong is
several meters below the level of the city, forming a cliff (Santoni 1997: 1; 2008:
90). This is very useful to avoid inundation in the wet season but it also looks like
the Mekong has destroyed a part of the city, a couple of hundred meters according
to the estimations (Rivolta 1997; Santoni 1997: 1; 2008: 90;). Hence we do not
know if there was a wall on the eastern side. Personally I believe not because of
the display of Sambor Prei Kuk we will be studying next.
52. 52
Figure IV.6: Aerial photography of Wat Phu. The walls are clearly visible at the bottom and on the left of the picture.
From Jacques 2004
The first feature of interest is the rivers. Only 3 of those flowing next to the city
are permanent, the Houay Phra Non, the Houay Sahoua and the Houay Khen
(Cucarzi & Zolese 1997: 27; Santoni 1997: 1). The others are dry during the dry
season, at least nowadays (Cucarzi & Zolese 1997: 27; Santoni 1997: 1). It seems
the rivers have been partly canalized (Rivolta 1997; Santoni 1997: 5); hence
having one going through the settlement was a will of the ones that established
here (Cucarzi & Zolese 1997: 28). This is very interesting if we remember the case
53. 53
of Angkor Borei that had chosen to have the Angkor Borei river going across the
site as well suggesting an unknown indigenous reason. In any case these canals
would have been able to bring the water from the mountain to the city. They clearly
were not used for transportation since the water only goes from the mountain to
the Mekong and the Mekong is already a major axis of communication. It seems
they were used solely for the purpose of bringing water.
Another interesting structure is the reservoirs. We find reservoirs of different
shapes in and around the city, some that tend to be nearly circular as the ponds
that existed in the Funan period, and some rectangular, very geometric, that we
did not see in Funan. They strongly remind the rectangular reservoirs from India of
course. And we can say they are very numerous in the area for a new introduction.
We do not know if they were all used together at the same time since no
systematic dating has been carried on but given the short span of the city and
where they are located a certain number of them had to be built at the same time
as the city. However it is their use that is the most relevant. A great number of
these, the religious reservoirs, are built in conjunction with religious structures to
their west side and inside the city. They are staircase-like made with flat rocks
probably coming from the Mekong (Cucarzi & Zolese 1997: 28; Rivolta 1997;
Souksavatdy 1997: 21). The religious structures are mainly Hindu but a double
stupa surrounded by a moat has been found4, hinting Buddhism was existing
along with Hinduism (Santoni & Hawixbrok 1998). These reservoirs are here for
sacred reasons only and there is no reason to think they contributed to the
network since they are rainfed. But another kind of these rectangular reservoirs,
located mainly outside the city, is connected to rivers or canals, often at deviations
(Rivolta 1997) (Fig. IV.8); indeed the area is flat and the watercourses tend to
meander, such systematic sharp angles could not have happened naturally. We
could of course suggest that they located the reservoirs at every sharp angle that
occurred naturally, however since the rivers have been partly channelled to go in
and around the city they are man-made and these angles are anthropogenic.
These reservoirs were clearly used with the intention of storing some water and
are not religious (Rivolta 1997). There are also a couple of them that are directly
4
This double stupa has been dated form the 7th
century, in the middle of the period of occupation of the
city. Why it was surrounded by a moat that is normally associated with a Hindu structure is unknown.
(Santoni & Hawixbrok 1998)
54. 54
connected to a watercourse by a short canal, without doubt in the aim of storing
water as well (Rivolta 1997). Hence they show there is an urge to store water that
we did not see in Funan. Why would they need for then? The environment is dryer
and the water precious in the area (Rivolta 1997) but so many reservoirs would not
be needed for domestic purposes only and they would not be outside the walls,
sometimes far from the city. Agriculture, in turn, seems more likely since these
reservoirs would allow a good complement of water for irrigated agriculture
(Rivolta 1997); indeed the whole area is suitable for agriculture within a 40 km
radius and as we have already said traces of irrigation seem to appear (Defente et
al. 1991: 194). They certainly could not distribute water to a big area (Rivolta
1997) but since they are scattered in the whole plain their number could
compensate. How the water was released from the reservoirs however is not
known but the Khmers are well known for building small temporary dams in the
canals or watercourses to force the water to inundate an area before destroying
these structures (Van Lière 1980). Hence we could think the water was released
perhaps in the very canals that brought the water in the reservoirs, and the canals
dammed in order to inundate the surrounding area. This is of course hypothetic
since we have yet to find such devices, however there seems to be traces of a
dam in Sambor Prei Kuk that could confirm they were already commonly in use at
that time. Unfortunately it has not been dated and the occupation extends beyond
the Chenla period in Sambor Prei Kuk so that it would be hasty to consider this
dam as being pre-Angkorian. Nonetheless, ethnology shows these practices are in
use today. In any case there was a clear will to store water that could only be
justified by an agricultural use.
56. 56
2. Sambor Prei Kuk
Our second case study concerning Chenla is Sambor Prei Kuk.
Sambor Prei Kuk is on the West bank of the Sen River which is a tributary of
the Tonle Sap. The area in itself is divided between a zone of temples and a city,
both separated by a river (Fig. IV.9). The temples have been first seriously
documented by Parmentier (1927) before being excavated much later by Groslier
in 1962. The city has been recognised from aerial photography by Goloubew
(1937: 655; 1938: 442) and serious excavation by the Sambor Prei Kuk
Conservation Project (SCP) led by the University of Waseda5 is now taking place
(Shimoda & Shimamoto 2012: 11).
The temples, on uplands to ensure they never get submerged, have been
constructed in the 7th century and renovated in the 10th century, in the Angkorian
period (Shimoda & Shimamoto 2012). They display several causeway linked to a
channel and rectangular reservoirs, barays, accompany the temples. These
barays are rainfed and the temples shelter Hindu symbols and statues, mainly
Shaivite (Shimoda & Shimamoto 2012). The barays were probably built at the
same time as the temples from the very beginning as they usually go together in
Hinduism as we have said. The causeways in front of the temples go straight to a
channel linking the O Krou Ke and the Sen River, suggesting for some
archaeologists the main entrance was through those causeways, reaching them
thanks to water transportation (Shimoda & Shimamoto 2012: 21). This also
supports the theory of a vast traditional water transportation history. From the point
of view of water management this section only contains rainfed barays for religious
purposes.
The city is to the West, separated from the temple area by a channel. It is
surrounded by shallow moats of a maximum depth of 1.5 meters that excludes
protection as a purpose and suggests religious intents. A 1 meter high dike is built
on the interior of those moats (Shimoda & Shimamoto 2012: 22). These moats are
nearly square-shaped with a 2 km side (Shimoda & Shimamoto 2012: 21) and
5
Complicating the finding of information for many of these articles are either written in the Bulletin of the
Graduate School of Waseda which is not easily accessible or in Japanese. Few articles on recent excavations
exist in English
57. 57
exist in the North, the West and the South. The channel makes the East side. It
has to be noticed that the channel is connected with the moats, directly in the
North and through a swamp on the South side. This reminds the technology from
Funan we have seen on the previous section. Nowadays the moats dry up in the
dry season but in the wet season part of the city is inundated (we will come back
on this issue) (Shimoda & Shimamoto 2012: 23). Inside the city nearly all the
buildings have been dated from the 7th century (Shimoda & Shimamoto 2012: 21-
24 & 67-69). Finally a possible dike-bridge has to be noticed in the middle of the
West moat (Shimoda & Shimamoto 2012: 23-24). It would seem this opening in
the moats has been practiced when the city has been built, dating it from the 7th
century. We can obviously see many striking similarities with Wat Phu, a square-
shaped city, water to the East and a separated temple area; thus this could appear
as the ideal model of city in Chenla.
And the comparisons with Wat Phu do not stop there: numerous Hindu
temples inside the city associated with rainfed barays, barays connected with
watercourses outside the city and many independent reservoirs outside the city
not necessarily connected to channels (Fig. IV.10). The latter feature indicates the
same desire to store water as in Wat Phu. The area is highly favourable to
agriculture nowadays and has high yields; this could once again suggest they
were used for agriculture. The finding of rice field boundaries around the city also
supports this idea (Shimoda 2010).
58. 58
Figure IV.8: The complex of Sambor Prei Kuk. The city is the walled area to the West while the temples are to the
East. From Shimoda 2012: Figure 2
59. 59
Figure IV.9: The city area of Sambor Prei Kuk. Shimoda 2012: Figure 5
As I have said before part of the city is now inundated in the wet season. I
believe there we have to mention the finding of the remains of a dam near the city
(Shimoda & Shimamoto 2012: 14). It is very possible part of the channel was
dammed during the wet season to regulate the flow of water. However this is only
hypothetical without dating of those remains and the ephemeral nature of these
installations in Cambodia make them hard to notice. There certainly was a way to
ensure the whole area could be used on a year-round basis.
In fact the only feature present in Wat Phu we do not see in Sambor Prei Kuk
is the use of canals, or at least they have not been reported. An explanation could
be found in the historical context: Bhavavarman I was said to have conquered the
territory and we could think that the establishment of the city might have been
hasty in order to ensure the possession of a capital; then we can hypothesize the
area has been chosen for its natural setting to avoid intensive investment of labour
and quick establishment.
60. 60
Sambor Prei Kuk bears many similarities with Wat Phu, ensuring some new
features such as a square-shaped design for the city, many Hindus temples and
an importance in the storage of water not seen in the Funan period. If we do not
see canals as in Funan or Wat Phu, the filling of the moats was certainly achieved
in a similar way as in Noen U-Loke or Angkor Borei.
Conclusion
Chenla seems to be the time of a whole new technological advance. The
synchronic apparition of Hindu temple and these new technologies could probably
be considered as a marker that change is happening. This is a whole cargo that
appears simultaneously in the city, one that strongly reminds Indian conceptions of
the city and the sacred. This new need for water could probably be linked to the
beginnings of irrigated agriculture as well and together with the apparition of this
Indian display could hint South Asia had a role in introducing this way of
cultivating. However some local technologies we have seen in Funan still exist,
such as the channelling of water, which reminds this system is also based on
some pre-existing indigenous knowledge.
D. Water management systems in India
As we have said in the first part of this chapter I will only review here some of
the water management systems that existed in India and I do not have the
pretention of addressing them all.
The site that has probably been the most studied in India is Sanchi, in Central
India. It has been known for a long time for its complex of hilltop Buddhist
monasteries but the water management system has only been studied since the
end of the 1990s (Shaw & Sutcliffe 2003b). It is a succession of hills with valleys
in-between suitable for agriculture. Traditionally wheat is grown there, however