Abstract In semi-arid Africa, environmental change continues to be debated, focusing particularly on the status and significance of land degradation and on local versus scientific knowledge claims. This paper suggests an approach for using different types of data sources, and for bringing together understandings of ecosystem dynamics and of people's interaction with the environment, and thereby achieving 'closure' in a highly contested terrain. The environmental and socio-economic history of a village in north-eastern Botswana is investigated using oral histories of villagers, aerial photographs (1964 and 1988), and official records. At first, descriptions of the environment appear confusing and contradictory, and at odds with official records. Rainfall is the focal point around which most explanations of environmental change turn. Here scientists find no long-term change, while villagers perceive a decline. Overstocking has been a main concern for policymakers throughout this century, while villagers do not see this as a problem. These, and ensuing contradictory perceptions, are partly resolved by identifying the main components linking the wider society to the local and to the environment, and by seeking a shared meaning between local and official versions, as well as between different local versions, of environmental change. We found, instead of contradiction, a striking convergence between, on the one hand, recent scientific understanding of the unpredictability and complexity of semi-arid dynamics and, on the other hand, the description and explanation of change offered by villagers themselves.