Network formation, learning and innovation in multi-stakeholder research projects : experiences with Adaptive Research and Learning Alliances in rice farming communities in Southeast Asia

Mounting pressure on research organizations to achieve sustainable development outcomes from research has pushed them to use multi-stakeholder approaches. Insights are missing however, on how these influence social, technical, and institutional change, as well as what outcomes emerge from these. The thesis is an examination of the enactment of multi-stakeholder approaches, questioning how and to what extent Adaptive Research (AR) and Learning Alliance (LA) approaches influence socio-technical innovation in rice farming communities. Four case studies of research and development projects that employed the approaches in rice farming communities were elaborated in this thesis. AR implementation in Indonesia (chapter 2), showed how AR fast-tracked technical adaptations and built upon the improvisational capacities of farmers. AR monitoring however, rendered invisible the adaptations required on the social aspect. Simultaneous social, technical, and institutional redesign was limited. A case of LA implemented at national level engaged a network that changed and expanded after three years to include diverse actors (chapter 3). There were points where implementation (mis)aligned with assumptions from project implementers and from conceptual literature of the LA approach. The network influenced change at community level by engaging small groups that made reconfigurations on the technologies and the social arrangements for these (chapter 4). A community-level LA in Myanmar was also found to stimulate a self-organized learning process towards innovation for flatbed dryer technology (chapter 5). A case where a project used AR only versus AR with LA in Myanmar (chapter 6), revealed differing networks, learning processes, and outcomes in terms of learning agenda. The involvement of a wider network resulted in a broader set of activities, which were initiatives outside the original plans of the project. The learning activities were not only about technologies but also included experimentations on supportive environment for access and use of the technologies. This thesis therefore demonstrates that project actors implement AR and LA approaches through a range of translations in multiple contexts. These imply varied interactions in different types of networks. Such interactions triggered varied learning processes and thus influenced different planned or emergent outcomes. Both approaches have potential to catalyze innovation in farming communities; however, outcomes on adoption numbers provide a caveat that these approaches are not silver bullets that guarantee technology adoption. Instead, implementation that facilitates effective learning processes, and monitoring that flags where projects could support emergent outcomes, can help implementers improve their contributions to development in farming communities.