Emergent Properties of Balinese Water Temple Networks: Coadaptation on a Rugged Fitness Landscape

For over a thousand years, generations of Balinese farmers have gradually transformed the landscape of their island, clearing forests, digging irrigation canals, and terracing hillsides to enable themselves and their descendants to grow irrigated rice. Paralleling the physical system of terraces and irrigation works, the Balinese have also constructed intricate networks ofshrines and temples dedicated to agricultural deities. Ecological modeling shows that water temple networks can have macroscopic effects on the topography of the adaptive landscape, and may be representative of a class of complex adaptive systems that have evolved to manage agroecosystems. N 1984, ERIC ALDEN SMITH PUBLISHED a devastating critique of the uses ofsystems ecology and simulation modeling in anthropology. While this article is in part a defense of these methods, we do not take issue with any of Smith's conclusions. Instead, we hope to demonstrate that systems models can serve a different heuristic purpose than the naive functionalist, energy-maximization or group-selection models skillfully demolished by Smith. In particular, we hope to show that simulation models are uniquely appropriate for addressing the issues of adaptation and determinism in the development of complex social systems like the water temples of Bali. But before we turn to the uses of simulation models, it may be useful to sketch out how our approach differs from those criticized by Smith. Although simulation models have always been a rarity in anthropology, they continue to be used extensively in biology as a tool to investigate complex interactive processes. For example, we recently served on the doctoral committee of a graduate student who was interested in the growth of algae in Antarctic sea ice, a major source of fixed carbon in the Antarctic Ocean. The student built a model to study the interactive effects of processes thought to influence the growth of the algae, such as temperature, nutrient flow, and available sunlight. The result was a system of differential equations that predicted, on purely theoretical grounds, variations in the growth of algae depending on the relationships among these causal factors. The model's predictions were then compared with observations, helping the student fine-tune his understanding of the mechanistic processes that drive the growth of the algae (Arrigo 1991). However, an obvious problem in extending this kind of analysis from biology to an