Fishermen, scientists, policy makers, and staff of environmental NGOs (ENGOs) have significantly different understandings of the processes that determine developments in fish stocks. These perception differences hinder the participatory debate on why fish stocks change and which management measures are effective. In this study, differences in causal reasoning about processes between fishermen, policy makers, ENGO-staff, and scientists were examined, regarding four case studies within the management of the fishery on North Sea plaice. First, it appeared that all parties, besides scientists, had difficulty reasoning about long-term effects because of comprehension problems with stock dynamics, and because of short-term economical interests. Second, there were differences in how parties deal with natural variation and interconnectedness of natural and anthropogenic influences. Stock assessment scientists work with single-species models, reducing complexity by using assumptions that rule out variation, in order to inform policy makers about the effect of one isolated management measure. Fishermen on the other hand, relying on information from their daily lives at sea, emphasize complexity and interconnectedness, and the impact of the ever-changing and unpredictable nature. ENGO-staff appeared reluctant to reason about single species and broaden the debate to the ecosystem-level, while emphasizing the effect of man. As a consequence of the diverging perceptions, much time in multi-stakeholder settings is lost on repetitive discussions, mainly on the relative importance of 'nature' versus 'man'. No wonder that policy makers feel lost, and experience processes as very complex. Concluding, to handle these perception differences, there is need for a directive process coordinator, and a more creative informative role for fisheries scientists. Together with all participants, they should map all expectancies and lines of reasoning at the beginning of the debate. This scheme can be relied on during subsequent meetings, in which perceptions can adequately be positioned.