Traditional irrigation in the dry zone of Sri Lanka

Sri Lanka's tanks

Sri Lanka is covered with a network of thousands of manmade lakes and ponds, known locally as 'tanks' (after 'tanque', the Portuguese word for 'reservoir'). Some are truly massive; many are thousands of years old: and almost all show a high degree of sophistication in their construction and design. Indeed, Sir James Emmerson Tennent, the 19th century historian, is fulsome in his admiration for those who built them.

In particular he marvels at the numerous channels which were dug underneath the bed of each lake in order to ensure that the flow of water was "constant and equal as long as any water remained in the tank". Frequently, he notes, those channels had to be cut through solid granite with the most rudimentary of tools:

"Their ruins present illustrations of determined perseverance, undeterred by the most discouraging of difficulties and unrelieved by the slightest appliance of ingenuity to diminish the toil of excavation." [1]

Today, the majority of the tanks which so impressed Tennent, have either totally or partially silted up. Nonetheless, numerous smaller tanks still survive (although many of these, too, are now partially silted up) and continue to provide the basis for irrigation agriculture in the dry zone of the island.

Large or small, the tanks are generally assumed to be the work of a centralised state bureaucracy - and hence it is argued (pace Wittfogel) that their silting up and subsequent abandonment can be explained by the breakdown of the state. That view, however, is not shared by Sir Edmund Leach, Professor of Anthropology at the University of Cambridge and a leading authority on irrigation agriculture in Sri Lanka. Thus, Leach argues that although the large tanks may have been the work of a bureaucracy, the small village tanks most decidedly were not. Indeed, his own research leads him to conclude that the Wittfogel hypothesis is quite inapplicable to Sri Lanka.

That conclusion is based on three critical considerations. Firstly, Leach argues that the primary function of the large tanks was not to provide water for agriculture - and certainly not for traditional subsistence agriculture. To be sure, the large tanks were used to irrigate farmland in the immediate vicinity of Sri Lanka's two historical capitals - first Anuradhapura and later (when that city was abandoned) Pollonaruwa - but it was not the peasants who benefited from such irrigation: the land was used principally to grow crops for urban consumption, to supply water to the capital or for purely ornamental purposes. [2]

By way of illustration, Leach cites the massive building programme undertaken by Parakrama Bahu, the megalomaniac king who reigned in Pollonaruwa from 1164 to 1197. Quite apart from building 101 temples and statues, Bahu also constructed numerous tanks. Those tanks, says Leach, were clearly built for the King's personal aggrandisement and had nothing whatsoever to do with satisfying the food needs of his subjects. "They are monuments not utilitarian structures."

Given the essentially ornamental nature of the large tanks, it follows - Leach's second point - that Sri Lanka's rural villages never actually depended upon the large tanks for their survival:

"When the central government was disrupted and the major tanks fell into disrepair, village life could carry on quite adequately. Each village still possessed its own small-scale irrigation system which was maintained by the villagers themselves." [3]

That last point brings us to Leach's third - and most telling - argument against those who would apply the Wittfogel thesis to Sri Lanka. Despite extensive research, he could find no evidence that a centralised bureaucracy ever even existed to run the country's irrigation works.

"On the contrary, the fact that, in the chronicles of the kings, the various monarchs receive praise for their munificence in repairing village tanks suggests that there was no routine procedure for carrying out such routine procedure on a national scale." [4]

That is not to say that repairs to the village tank were neglected. Far from it. The crucial point, however, is that the necessary maintenance work was organised by the villagers themselves: there was never a centralised bureaucracy to direct such work or to ensure that it was carried out. As Leach writes,

"From time immemorial, normal repair work to the village tanks has been the ordinary work of ordinary people. Major repairs and new constructions were traditionally undertaken by a specialised caste group of Tamil labourers - the Kulankatti - but these people worked for the villagers on direct contract: they were not employees of the state."

The running of the village irrigation system was thus firmly in the hands of the local community. Indeed, according to Leach,

"it is only since about 1860 that a centralised Irrigation Department has had the right to interfere on matters relating to the maintenance and use of village tanks." [5]

Traditionally, much of the maintenance work on the tanks was carried out during the Rajakariya - the forty-day period when every Singhalese villager was required to work free for the King. The Rajakariya should not, however, be seen as constituting a state-run maintenance programme. The villagers were not indentured labourers (a point which was lost on the British who abolished Rajakariya service as a distasteful relic of feudalism) nor were they employed by the state. On the contrary, the work was organised at the local level. Moreover, the villagers had a considerable say in the work they undertook:

"indeed, on one occasion, they refused to dig an artificial lake outside the King's palace in Kandy on the grounds that it was an ornamental showpiece and should not, therefore, be built with Rajakariya labour." [6]

We have, then, a very different picture from that painted by historians of the Wittfogel school. The latter, it will be remembered, sought to explain the rise of the state in early 'hydraulic' societies by the need for a centralised bureaucracy to run their irrigation works. Without such a bureaucracy, it was argued, irrigation agriculture could not have been practised.

The evidence from Sri Lanka, however, firmly refutes that argument: not only did villages run their own irrigation systems quite independently of the state but - and this is critical - they continued to do so even after the state effectively collapsed. (The more that one learns about traditional irrigation systems, the less the Wittfogel thesis seems justified. Sutton, for instance, has pointed out that all the societies which practise irrigation in East Africa are democratic to the point of not even having a chief. Instead, their system is run by a council of elders which is independent of the main political structures.)

If, therefore, there is any connection between the practice of irrigation agriculture and the rise of the state, the explanation for it must lie elsewhere than in the need for a bureaucracy to run the irrigation works. As Leach asks,

"Could it be that the sociological explanation of why so many of the ancient societies were 'hydraulic' is that, in a wide variety of circumstances, hydraulic society lends itself very readily to the development of specialised labour on a non-monetary basis?" [7]

Even that interpretation, however, has its problems. For 'specialised' work to be undertaken 'on a non-monetary basis' implies - at the very least - the existence of a cohesive community bound together by reciprocal rights and duties. It also implies a common cultural pattern, adapted exclusively to the practice of irrigation agriculture. Such a community is the very antithesis of a state. If, therefore, a state were to arise from it - through the increasing division of labour - then one would have to posit the breakdown of the very communal ties which had previously held it together.

The importance of those communal ties has been stressed by almost all the authors who have studied the traditional practice of irrigation in Sri Lanka. In that respect, it is worth quoting Sir James Emmerson Tennent.

"Cultivation, as it existed in the north of Ceylon," he writes, "could only be carried on by the combined labour of the whole local community, applied in the first instance to collect and secure the requisite (water) supply for irrigation and afterwards to distribute it to the rice lands which were tilled by the united exertions of the inhabitants, among whom the crop was divided in due proportions. So indispensable were concord and union in such operations that injunctions for their maintenance were sometimes engraved on the rocks, as an imperishable exhortation to forbearance and harmony." [8]

Significantly, Tennent rejects the suggestion that Sri Lanka's irrigation works broke down as a result of faulty construction and, in particular, the absence of spillways for draining off surplus water during the rainy season.

"For upward of fifteen centuries the reservoirs, when duly attended to, successfully defied all the dangers to be apprehended from inundation." Besides, "vast numbers of these tanks, though utterly deserted, remain in this respect, almost uninjured to the present day." [9]

Instead, insists Tennent, the destruction and final abandonment of the tanks should be seen as the inevitable outcome of social decay - and in particular, "the disruption of the local communities by whom they were so long maintained". With that disruption came an end to that 'concord and union' which Tennent held to be so critical to the running of the irrigation works. The consequences were undoubtedly disastrous:

"The ruin of a reservoir when neglected and permitted to fall into decay was speedy and inevitable; and as the destruction of the village tank involved the flight of all dependent upon it, the water, once permitted to escape, carried pestilence and miasma over the plains they had previously covered with plenty. After such a calamity, any partial return of the villagers, even where it was not prevented by the dread of malaria, would have been impracticable, for the obvious reason, that where the whole combined labour of the community was not more than sufficient to carry on the work of conservancy and cultivation, the diminished force of the few would have been utterly unavailing, either to effect the reparation of the watercourse, or to restore the system on which the culture of rice depends. Thus the process of decay, instead of a gradual decline as in other countries, became sudden and utter desolation in Ceylon." [10]

It is a warning which the government of modern Sri Lanka would have done well to ponder upon before it embarked on the giant Mahaweli scheme.

The importance of the tanks

The traditional Ceylonese village was dominated by three features: the temple (dagoba), the tank (wewa) and the paddy field (ketha). Such is the importance of the tank that, according to the late Upalli Senanayake,

"one could not imagine a village in the dry zone without a tank any more than one could imagine it without a temple or a rice paddy." [11]

Indeed, the tank was judged to be so vital to village life that the term 'wewa' was frequently used synonymously with the term 'gama' or village. As R. H. Brohier, one of the foremost authorities on Sri Lanka's ancient irrigation system, puts it:

"People say they belong to Siyambalagaha-Wewa, 'the Tamarind tree tank', and not to Siyambalagaha-gama, 'the Tamarind tree village'." [12]

Ideally, several different types of tank were built - some of which had nothing to do with irrigation per se but all of which had a critical role to play in the practice of irrigation agriculture. It was, for example, traditional to build a forest tank in the jungle above the village. That tank, however, was not used to irrigate land: on the contrary, its express purpose was to provide water to wild animals and, hence, to reduce the likelihood that they would descend into the paddy fields and destroy the crops in the search for water. Other tanks included:

• The mountain tank, which was built to provide water for 'chena' or slash-and-burn agriculture - a vernacular form of farming now frowned upon (if not actually discouraged) by the authorities.
• The erosion control tank, or 'pota wetiye', which was so designed that any silt was deposited in it before entering the main water storage tanks. Several erosion control tanks were associated with each village irrigation system. All were built in such a way that they could easily be de-silted.
• The storage tank, of which, traditionally, there were two - one being used whilst the other was being repaired. For that reason, such tanks were known as 'twin tanks'.
• The village tank, of which there was one for each village that depended upon a particular irrigation system. All such tanks were connected by canals to the twin tanks. [13]

Irrigation as a way of life

Paddy growing was not an occupation: it was a way of life, closely interwoven with other social activities. Each stage in the agricultural cycle - from weeding, to ploughing, to transplanting the paddy and, finally, to harvesting it - was accompanied by special ceremonies involving song, music and dance. Indeed, those traditional dances which still survive clearly originated in such ceremonies: they are based on rhythmic movements which visibly symbolise reaping, ploughing and digging.

Significantly, it was the priest who initiated the most important agricultural activities. When the time was deemed auspicious for ploughing, for instance, the temple bell would ring and the whole village would stream out into the fields. The King himself would participate in the ploughing ceremonies that took place near the ancient capitals.

As in all peasant societies, agriculture was very much a family affair. Each member of the family, including the children, had specific responsibilities. One child's job, for instance, was to drive away any marauding monkeys from the paddy fields: another's was to look after the cattle and water buffaloes. One or two children would help their father in the fields: the rest would help their mother to harvest firewood, to prepare food and to milk the cows and buffaloes - the girls being specifically responsible for weeding and for making mats.

So, too, there was a tradition of mutual help - 'attama' - within the village; neighbours could thus be relied upon to help with pressing, day-to-day chores and, more important still, with the more onerous agricultural tasks. [14]

The sustainability of the traditional system

Leach points to the great stability of the village and its irrigation system. Whereas governments rose and fell, the village and its tanks remained the same for thousands of years.

"Under Ceylon dry zone conditions, once a village and its irrigation tank have been constructed, it is there forever and since the irrigation area must always remain the same size, the population of the village itself can only vary between very narrow limits." [15]

As in other traditional societies, the Singhalese regarded their institutions as permanent. Thus, old inscriptions recording donations of land to various temples, stress that they would be valid 'ira handa pavathina thuru' - "so long as the sun and the moon are there." [16]

Under the traditional system, that faith in the future was well founded. As with other peasant societies, the agricultural system was geared to minimising risks rather than to maximising yields. Thus, in order to guard against the upheavals of drought, pests, floods and similar agricultural disasters, farmers would plant a wide variety of different strains of the same crop.

Indeed, Mudyanse Tenakoon - a local farmer and philosopher whom we interviewed in 1982 - recalls that over 280 different varieties of rice were in common usage during his youth; today, however, only 15 survive. Each variety had different characteristics and each was capable of surviving in conditions which would threaten the other varieties. [17] (According to C. Dreiberg [Superintendent of School Gardens], quoted by C. Wright in Glimpses of Ceylon, 1874, three to four hundred varieties of rice were once cultivated.)

So, too, custom prohibited the construction of permanent buildings on prime agricultural land. Only the King and the priests, for instance, were entitled to build brick and tiled houses: everybody else lived in mud huts. That custom was founded on sound ecological principles. Brick houses do not break down when they collapse: instead they sterilise the soil. (Here one might note the terrible damage done by brick-works in Egypt and India. In both countries, vast areas of land have been dug up to a depth of several feet in order to provide earth for bricks.) Mud houses, on the other hand, quickly return to the soil, thus providing valuable organic matter for the fields.. Significantly, Tenakoon still refuses to this day to live in a brick house.

Most important of all, however, the civilisations of ancient Sri Lanka protected their forests. If we are to believe Robert Knox - the sixteenth-century freebooter who was shipwrecked on the island and subsequently spent 15 years as a captive of the King of Kandy - it was religious beliefs, rather than government ordinances, which prevented the forests from being cut down. According to Knox, the Singhalese believed that when their ancestors first invaded Sri Lanka, the hostile spirits of the island's original inhabitants sought refuge in the jungle-clad hills of the highlands. To enter those jungles was thus considered singularly unwise.

Whatever the truth of that explanation, it is certain that the protection of the highland forests played a vital role in ensuring the sustainability of agriculture in the dry zone. Indeed, it might even be argued that adequate forest cover in the highlands was the sine qua non of a healthy irrigation system. Not only did the mountain jungles intercept and store the monsoon rains, but they also regulated the flow of water to the island's rivers. Without them, a perennial and certain supply of water is by no means certain. The point is well made by R. H. Brohier:

"The central mountain region ... consists of hill piled over hill, and mountain range over mountain range, on a succession of ledges of great extent, at various elevations....

Ravines of various depths which form conduits for the mountain torrents lead from each one of these mountain ledges to the other. On the comparatively level tracts they form large swamps. All but two of the larger rivers of Ceylon collect their waters from these swamps in the mountain ledges. Rushing down from the mountains and following the grain of the country between the foot-hills, they meander sluggishly over the rest of their course which lies in the maritime plains. It is easy, therefore, to realise that the first principle behind the whole system of water storage and carriage which, in ancient times helped in irrigating the low-country, is centred on the natural reservoirs up in the mountains.

A thousand years ago, as much as today, these centrally situated mountain heights of the island no doubt served to intercept the monsoon currents: but, whereas in the centuries gone, these primaeval forests ... helped to condense the vapour-laden clouds, conditions today tend to dissipate them. We may, therefore, assume, leaving little room to doubt, that in the past when large river-fed works of irrigation functioned over the plains, there were influences which induced a much more abundant and liberal rainfall over the mountain zone.

Then, again, the wooded slopes, with foliage acting as a parasol to the ground, served to break the force of the rainfall, to retain the surface soil and to help the ground to absorb some of the moisture. In this last respect, it exercises an important influence in forming sub-soil springs which afforded the rivers a means of maintaining a perennial flow." [18]

After the fall of the ancient kingdoms, the seat of Singhalese civilisation moved to Kandy in the uplands. It is probable that a measure of deforestation was necessary to accommodate the city and its surrounding agricultural areas. Deforestation began in earnest, however, with the arrival of the British and, in particular, with the setting up of coffee (and later tea) plantations in the mountains. Even so, at the time of independence, Sri Lanka was still 40 percent forested. Since then, whole areas of jungle have been cut down in order to earn foreign currency from the sale of timber: today, forests cover a mere 4 or 5 percent of the island.

The ecological consequences of such deforestation have proved devastating. Indeed, it would be no exaggeration to say that the virtual disappearance of the country's jungle cover has cast serious doubts upon the viability of agriculture in many areas of the island. "Today, streams which were once perennial expose dry and stony beds", reports Brohier.

"As a result of the short, swift descent from the mountain zone, the monsoon torrents, which are no longer restrained, sweep down the rivers carrying debris, earth and sand. This tendency to sudden flooding in the plains has gradually increased the marshes. It has changed the physiography of many a region which was described in ancient history as proverbially fertile. The effect of denudation has, in fact, been transmitted throughout the entire course of the rivers. This is indicated by the shoals and sand banks and sand-bars at the river mouths." [19]

Brohier goes on to comment:

"It will be safe to assume that this analysis of altered conditions will come as a revelation to many who proclaim that every ancient tank and channel system is capable, if restored, of fertilising vast areas that are barren and unproductive." [20]

Certainly, it will come as a revelation to the unfortunate bureaucrats who have been called upon to run the massive complex of dams and irrigation works at present being constructed on the Mahaweli River.

Land tenure: traditional vs. modern

Under British Colonial rule, all land whose ownership could not be firmly established was confiscated, designated "crown land" and then sold off in plots of five acres. Land thus became a marketable commodity in a country where, previously, the only items which it had been permitted to trade were salt, salted fish and clothing. (And even then the profit from such trade was limited by law to a quarter of the value of the produce sold.)

Five acres is a large tract of land for a villager in Sri Lanka: poor peasants were thus effectively excluded from buying the new plots and, inevitably, the pattern of land-tenure changed dramatically. Those changes are eloquently described by Edmund Leach in Pul Eliya, his classic 1953 study of a dry zone village. [21] Thus, he notes, land in Pul Eliya is now divided into two categories: the 'Old Field' (the 40 or so acres which had escaped confiscation) and the 'New Land', (that land, in other words, which had been sold by the colonial authorities.)

The Old field is divided into two sections: the 'upper field' and the 'lower field'.. Those sections, in turn, are divided into more than one hundred strips of land which are farmed by different families. The distribution of those holdings is, as it always was, strictly egalitarian: thus, those

"who own land in the lowest and least advantageous portion of the upper fields also own land in the highest and most advantageous portion of the lower field" - the latter being the area with first access to irrigation water. [22]

By contrast, land-holdings in the New Lands are distributed according to the strict laws of the market. Here, the rich control the best land and the poor make do with the marginal land - a situation which, says Leach, "is bound to have very drastic implications for the village considered as a social entity". [23]

Despite the pressures of the market, however, those who farm the new lands have done their utmost to reproduce the fragmented pattern of land holdings which characterises the Old Field system. For instance, when the government originally put up the land for sale, many villagers formed themselves into syndicates to buy land and then divided it up amongst their members. Large plots were thus broken down into smaller, individual holdings.

Such syndicates, however, have brought their own problems to the management of the New lands. Thus, as Leach points out,

"the most likely group of people to operate and maintain an irrigation system is a group of kinsmen who are capable of co-operating in a way that people unrelated by family bonds cannot." [24]

No such bonds exist between the members of many of the syndicates which have bought the New Lands: consequently, co-operation between them has proved difficult to achieve. Clearly, it is a problem of which the villagers themselves are well aware: indeed, Leach notes that they have sought "to re-establish the old form of co-operation among the owners of the New Land by means of kinship alliances". In some cases, kinsmen have themselves formed syndicates in order to buy plots in the New Land and work on common irrigation channels.

Finally, the New Lands have been dogged by problems over the allocation of water. In the Old Field, water is allocated on the basis of a time-honoured rota system which ensures that everyone receives their fair share at the right time. The system (which, in Leach's view, is so highly 'traditionalised' that it is impossible to change) is supervised by the vel vigane or 'irrigation headman', who also overseas the maintenance of the tanks and channels. (In earlier times, it had been the vel vigane who organised the Rajakeraya.

According to Tenakoon and Senanayake, the vel vigane was originally elected by the villages to supervise their irrigation works.. Leach, however, regards the post as a creation of the British. Although the allocation of water in the New Lands is still assured by the vel vigane, the latter is no longer subject to the social constraints that operate in The Old Field - nor is the traditional method of allocating water used in the New Lands. As a result, writes Leach, the vel vigane and his friends now have "a dominating economic position in the village". Inevitably, the allocation of water in the New lands has become both inequitable and arbitrary. [25]

That problem is compounded by the use of three irrigation channels in the New Lands, each of which is separately owned and managed. Thus, as Leach notes, "the control of the channels corresponds very closely to the general pattern of factionalism" which now divides the village.

By contrast, the Old Field has just one irrigation channel which is operated and maintained by the whole community. Indeed, it is a measure of the failure of the new lands that they suffer from chronic water shortages and are only able to grow one crop a year, whereas - under the old Field System - two crops were grown each year. In part, that failure can be explained by the increased acreage under cultivation: there simply is not enough water to irrigate the whole area twice a year. Far more fundamental, however, is the factionalism and divisiveness which plague the New Lands. In an agricultural system which relies on co-operation, such strife is surely a recipe for disaster.

The growth of the state and the breakdown of the traditional system

With the development of the market economy there also came a rapid expansion in the powers of the state. In fact, one of the first steps taken by the British colonial administration was to strip the responsibility for maintaining the tanks from the villagers and to place it in the hands of a central irrigation department. The results were catastrophic. As one official was to put it to a select committee of the House of Commons in 1849: "What was everybody's business has became nobody's business". [26]

Today, the irrigation system of most dry zone villages is crude in comparison to that which existed under the ancient civilisations. Rare indeed is the village which still has its full complement of operational tanks. Although, recently, a certain amount of work has been done to de-silt the larger tanks, the government has not seen fit to include the smaller tanks in the programme.

Nor should that surprise us. To the Irrigation Department, the smaller tanks are arcane relics of the past: their use cannot possibly be justified on the basis of conventional cost-benefit analysis and, therefore, in the eyes of the Irrigation Department, they are simply redundant. How, after all, could the government sanction the maintenance (let alone the construction) of a tank purely for the benefit of wild animals? On what economic grounds could it conceivably justify the building of two storage tanks when to do so would mean doubling the cost of both construction and maintenance? And who would pay for the bulldozers and other capital-intensive equipment to de-silt the smaller tanks?

The sophistication of the traditional irrigation has thus been sacrificed in the interests of economic expediency. With it has gone a whole way of life: the tradition of mutual help has all but disappeared and many essential agricultural tasks - in particular weeding - are no longer properly carried out. (Senanayake was so shocked by the state of the country's paddy fields in the early sixties, that he persuaded his cousin Dudley, who was then Prime Minister, to allow him to mobilise several hundred thousand school children to do the weeding for free. The scheme ran for several years.)

It would seem that history is repeating itself. Just as the irrigation system of the ancient civilisations fell apart when the communities which maintained it were disrupted, so today we have a new wave of abandonment. It remains to be seen whether Colombo will go the same way as Anuradhapura and Pollonaruwa.

References