His current work concerns the role of learning processes in local development projects, and the challenges of combining these processes with the dimensions of partnership, accountability and the various differences shaping international aid relationships (including cultural ones). A key aspect of this is how, within these dimensions, the evidence of projects’ success is established as adequate, and how it is translated into the indication that learning and change are needed, for instance, over the course of jointly organised evaluations. Learning in the practice of development projects involves determining what went wrong and what should be improved, as well as why, and how. This implies ideas of better practice, as well as concepts of evidence (of success or failure, of room for improvement, or of producing evidence). More often than not, these notions originate on the financing side of cooperation. This is a problem in particular when cooperation is framed as “helping people help themselves”, emphasising local ownership and partnership “on eye’s level” despite the power differences between international financiers and local grass-root actors. Attempts to organise learning in development cooperation could easily be conceived of as a minefield, with looming accusations of neo-colonial “smart-arsing” (cf. Lepenies 2009) through the backdoor. At the same time, evaluations often do succeed in the sense that a jointly drawn picture results, including the open identification of deficits, based on which changes in the cooperation are agreed upon. In case this happens, the question is: How? How does cooperation in evaluation and other learning processes work, despite the aforementioned tensions? How do parties cooperate in bringing together the production of evidence of success, which is needed for the accountability dimension of evaluation (i.e. for the license to continue), with the identification of deficits needed for the learning dimension?