Tuesday, April 20, 2010

Translating local knowledge into state-legible science

James Scott’s Seeing Like a State (about which I've blogged previously) talks about how the state imposes standards in order to make features legible, countable, regulatable, and taxable. J. Stephen Lansing’s Perfect Order: Recognizing Complexity in Bali describes a case where the reverse happened. When Bali tried to impose a top-down system of scientifically designed order--a system of water management--on Balinese rice farmers, in the name of modernization in the early 1970s, the result was a brief increase in productivity followed by disaster. Rather than lead to more efficient use of water and continued improved crop yields, it produced pest outbreaks which destroyed crops. An investment of $55 million in Romijn gates to control water flow in irrigation canals had the opposite of the intended effect. Farmers removed the gates or lifted them out of the water and left them to rust, upsetting the consultants and officials behind the project. Pesticides delivered to farmers resulted in brown leafhoppers becoming resistant to pesticides, and supplied fertilizers washed into the rivers and killed coral reefs at the mouths of the rivers.

Lansing was part of a team sponsored by the National Science Foundation in 1983 that evaluated the Balinese farmers’ traditional water management system to understand how it worked. The farmers of each village belong to subaks, or organizations that manage rice terraces and irrigation systems, which are referred to in Balinese writings going back at least a thousand years. Lansing notes that “Between them, the village and subak assemblies govern most aspects of a farmer’s social, economic, and spiritual life.”

Lansing’s team found that the Balinese system of water temples, religious ritual, and irrigation managed by the subaks would synchronize fallow periods of contiguous segments of terraces, so that long segments could be kept flooded after harvest, killing pests by depriving them of habitat. But their attempt and that of the farmers to persuade the government to allow the traditional system to continue fell upon deaf ears, and the modernization scheme continued to be pushed.

In 1987, Lansing worked with James Kremer to develop a computer model of the Balinese water temple system, and ran a simulation using historical rainfall data. This translation of the traditional system into scientific explanation showed that the traditional system was more effective than the modernized system, and government officials were persuaded to allow and encourage a return to the traditional system.

The Balinese system of farming is an example of how local knowledge can develop and become embedded in a “premodern” society by mechanisms other than conscious and intentional scientific investigation (in this case, probably more like a form of evolution), and be invisible to the state until it is specifically studied. It’s also a case where the religious aspects of the traditional system may have contributed to its dismissal by the modern experts.

What I find of particular interest here is to what extent the local knowledge was simply embedded into the practices, and not known by any of the participants--were they just doing what they've "always" done (with practices that have evolved over the last 1,000 years), in a circumstance where the system as a whole "knows," but no individual had an understanding until Lansing and Kremer built and tested a model of what they were doing?

[A slightly different version of the above was written for my Human and Social Dimensions of Science and Technology core seminar. Thanks to Brenda T. for her comments.  More on Lansing's work in Bali may be found online here.]

1 comment:

Monado said...

I can imagine some very fruitful research projects to extract procedures from practices and see what mechanisms are embedded in the practices. We could carry modern workflow analysis from the factory and office to the farm and forest and perhaps preserve some of that embedded knowledge before it's lost.