The universal application of the T&V model of agricultural extension in more than 50 countries is one of agricultural development’s best known failures. The approach worked well in places where it was originally developed, but proved inappropriate almost everywhere else. In this report Rasheed Sulaiman V. and Andy Hall worry that an apparently successful extension innovation piloted in India is set to suffer a similar fate.

The question of how agricultural research can best be used for developmental purposes is a topic of some debate in developmental circles. The idea that this is simply a question of better transfer of ideas from research to farmers has been largely discredited. Agricultural innovation is a process that takes a multitude of different forms, and, within this process, agricultural research and expertise are mobilised at different points in time for different purposes. This paper uses two key analytical principles in order to find how research is actually put into use.

Innovation platforms (IPs) are a way of organizing multistakeholder interactions, marshalling ideas, people and resources to address challenges and opportunities embedded in complex settings. The approach has its roots in theories of complexity, the concept of innovation systems and practices of participatory action research. IPs have been widely adopted across Africa and beyond in recent years as a “must have” tool in a range of “for development” modes of agricultural research.

This paper explores the use of complex adaptive systems theory in development policy analysis using a case study drawn from recent events in Uganda. It documents the changes that took place in the farming system in Soroti district during an outbreak of African cassava mosaic virus disease (ACMVD) and the subsequent decline in cassava production — the main staple food in the area. Resultant adaptation impacts are analysed across cropping, biological, economic and social systems each of which operate as an interlinked sub-system.