There's method in some messes

Andy Ho, Straits Times 19 Feb 09;

SINGAPORE is a dull city, says Mr William Lim. The architectural theorist says we disvalue our Geylangs and Little Indias. So we tear down the old to build new edifices without recognising the charm of these 'spaces of indeterminacy'. Despite their seeming messiness, 'there is a very unstructured order' about these places, he argues.

Spoken like a fan of complexity theory - the subject of a conference that the Nanyang Technological University co-organised last week. The idea is that in systems made up of interacting parts that can adapt to the results of their interactions - complex adaptive systems, they are called - lower-order interactions following simple rules can result in structured, higher-order patterns.

Like cells, tissues and organs that combine to form life, say. Or the millions who make up cities. Or Balinese rice terraces, which have been flourishing for 1,000 years.

At the conference, University of Arizona anthropologist Stephen Lansing sketched a case study of how self-organising, self-repairing order emerged from unplanned interactions among Balinese rice farmers.

Bali slopes down from several volcanoes in the island's centre. The slopes are dissected by rivers along which, for a millennium, small groups of farmers have held regular meetings in 'water temples' to manage their irrigation systems.

These temples serve to link places and social groups. Water management groups called subaks hold regular meetings in them to integrate irrigation decisions.

Networks of subaks form the 'congregations' of higher level water temples, which together add up to a top level water temple at the watershed level. Every year, representatives make a pilgrimage to the top to bring down holy water to mix into their subak level water sources.

For a millennium, subaks have made decisions collectively, consensually and democratically, with no one asserting top- down oversight. This is amazing considering that Balinese social life is otherwise dominated by a caste structure evident even in the way they speak today.

Contiguous subaks weigh how much water each gets, when to plant, when to flood the fields, and so on. The rituals at these water temples remind everyone of the totality of the whole system.

While each subak cares only about its own farming problems, an overall solution emerges that optimises irrigation flows for all parties. In this way, labour-intensive, complexly connected, irrigated rice cultivation has been sustained in this mountainous island of deep valleys. With little flat arable land, Bali should see dispersed, small scale, highland dry farming instead.

This hydraulic system brings water from the mountain tops down to the padi terraces along the slopes quite equitably, with no overall administrator. How did the subaks self-organise into a yield-enhancing complex adaptive system?

After all, upstream folks have their hands on the spigot. So why be nice to downstreamers? The answer: pest control. If everyone planted at the same time, there would be only one brief window of opportunity for their specific species of pests. Pests were thus starved the rest of the time. If upstreamers refused to open the spigot enough, downstreamers could refuse to synchronise their planting. Then, pests would move upstream to devour their rice crops. Thus the 'generosity'.

Still, how did such order get started if no king had set down the rules aeons ago? Working with complexity scientists at the Santa Fe Institute which co-organised the NTU conference, Prof Lansing built an agent-based model of 172 subaks planting at random times while observing and imitating successful neighbours.

The computer model was run to see if such an approach would increase crop yields. Lo and behold, it self-organised within 10 years into a system that mapped onto the actual one in Bali. In the model, yields were optimised when subaks synchronised what they did with their successful neighbours.

What if there were top-down irrigation management instead? In the 1960s, Western experts recommended five-year development plans for the Third World. The subak system conflicted with such plans. In 1971, the Indonesian government - egged on by Western experts during the Green Revolution - dictated the use of high-yield strains, fertilisers and pesticides as well as planting as often as possible. Within two years, pests had drastically reduced crop yield. Predictably, pesticide use escalated but to no avail.

By the 1980s, the government had ended the 'plant often' idea and reverted to synchronised planting. Pesticides were no longer needed while the traditional approach provided enough nutrients as rain water leached minerals from the rich volcanic soil into rivers and springs. Yields increased again.

A simple rule - 'synchronise your planting' - led to pest control and equitable resource sharing. This indicates that simple rules can lead to the bottom-up emergence of complex adaptive systems. By contrast, the 1971 decision to dictate strains and so on, showed what can happen when self-organising principles are ripped out.

So Mr Lim may well have a point. If we left the seemingly 'messy' parts of Singapore alone, the system could take just 10 years to self-organise into creative ordered disorder.