The persistence of a reversed reading of Kissidougou's landscape over a century seems hard to credit. In this chapter we explore some explanations for its endurance, examining how degradation visions have succeeded in excluding other environmental perspectives, including those of local inhabitants, and thus why the vegetation changes which have occurred have been rendered invisible.

The degradation position, we will argue, has been sustained not on the basis of ignorance, but through the continual production of supportive knowledge. Those who are convinced of deforestation and savannisation do not lack ‘evidence’ to draw on in support of their convictions. At stake here is partly the inheritance and transfer of scientific ideas and supportive theories. But while tracing the genealogy of landscape readings through individuals and institutions is useful, it is also inadequate. It overlooks how successive generations of observers repeatedly ‘rediscover’ readings for themselves, within common sets of intellectual structures and social relations. The persistent conviction of savannisation, we suggest, owes a great deal to this process of continual derivation.

In this production of knowledge, the concepts and methodological structures of ecological sciences have been centrally important, but the relations of production extend beyond ‘science’ itself. This chapter therefore explores the social, political and financial conditions which have mutually supported each other in upholding this powerful backwards reading of history, and the ways that local political processes have been incorporated into it.