Farmers in Nepal have previously been credited with a sophisticated understanding of the fodder value of different tree species and how this varies seasonally, much of which is encapsulated in two independent classification systems that they use to distinguish fodders of different utility. These farmer classifications have also been found to correspond to biologically pertinent attributes of fodder such as protein supply and overall dry matter digestibility. Here, a model derived from these correspondences and incorporating fuzzy sets, is used to explore the utility of combining the qualitative knowledge of farmers about a wide range of tree species with quantitative scientific data and understanding. Although a simple example of what may be possible using these methods, the result is a potentially powerful tool for interpreting decision making by farmers in a complex domain. This type of approach might be widely applied to make more complete use of qualitative information garnered from local people in the process of participatory rural appraisal.