Abstract
Swidden (aka 'slash-and-burn') agriculture is a prototypical-coupled human and natural system. Despite this, mathematical models integrating its social and ecological dynamics are rare. Here, we use complex adaptive systems theory to develop a model where individuals rely on labour exchange driven by reciprocity and normative reasoning that can lead to sanctions. Our results identify three emergent regimes: low-intensity swidden, sustainable high-intensity swidden that maximizes ecosystem services and harvest returns, and deforestation. We show that sustainable high-intensity swidden evolves if labour reciprocity and normative reasoning are balanced: helping behaviour should be significantly conditioned by normative reasoning to prevent over-harvesting, while reciprocity is necessary to prevent excessive sanctioning. We find that the sustainable high-intensity swidden regime is robust to changes in group size, is resilient to environmental shocks, can evolve under various models of forest ecology and is most productive for both forests and farmers when the balance of labour reciprocity and normative reasoning results in an intermediate scale of forest disturbance. Overall, we illuminate the importance of Indigenous social norms and customary practices related to swidden labour in maintaining sustainable and intensive swidden agriculture.