The main question addressed by this article is how to adapt the responsive evaluation (RE) approach to an intervention context characterized by repetition of ineffective interventions, ambiguous intervention action theories among stakeholders, and high complexity. The context is Grand-Popo, a fishing municipality located on Benin's southwest Atlantic coast. The fishery management interventionists and the fishing communities in the municipality all espoused concern for the sustainable improvement of fishing actors’ livelihood conditions, but differed about the reasons for this livelihood impairment, and about what should be done, when, where, and by whom. Given this ambiguity, we identified RE as a promising action research approach to facilitate dialogue and mutual learning, and consequently to improve stakeholders’ ability to resolve problems. However, this approach seems to have some shortcomings in the Grand-Popo context, regarding the repetitive ineffectiveness of interventions, high complexity, and uncertainty. Therefore, based on our empirical study, we add three dimensions to the existing RE framework: historical analysis to deal with routine interventions, exploration and discussion of incongruities of action theories to trigger double-loop learning, and system analysis to deal with complexity and uncertainty. This article does not intend to address the implications or impact of this adapted RE framework. Instead, we suggest some criteria and indicators for evaluating whether the proposed amended RE approach has assisted in resolving the fishery problems in Grand-Popo after the approach has been applied.
