The notion that pastoralists are irrational managers due to strong adherence to tradition and culture is still common in livestock production sciences. Researchers and development practitioners tend to fall back on this notion when target groups do not adopt their proposed innovations without any obvious reason. It is however difficult to identify innovations that fit into resource-poor systems, and often this lack of fit is the reason why innovations are not taken up. Understanding why pastoralists do what they do, and learning about the constraints they face when regulating production processes, is a prerequisite for identifying viable improvement possibilities.

This study focuses on understanding the reasoning behind pastoralists’ actions. The knowledge underlying pastoralists’ regulation of production processes is analysed with a method based on second-order cybernetics that offers insight in the system view of the actors in the system. It uses feedback and control principles of cybernetics to understand human control in human activity systems. For analysis purposes, a control loop model is proposed to systematically assess, for each management practice, the livestock keepers’ observations and the rules upon which they base their actions.

This cybernetic knowledge analysis is illustrated using the example of milk offtake management in a pastoral camel production system in northern Kenya. The example is chosen because lack of knowledge on the complex milk offtake management led to the widespread preconception of poor pastoral management as the cause for high calf mortality.

The cybernetic knowledge analysis reveals what livestock keepers consider as information in their production process, as well as the rules that inform their actions. The analysis enables to distinguish between management practices that serve as: (i) routine control; (ii) problem-solving control; or (iii) selection. This information is useful in transdisciplinary studies that involve local actors and other stakeholders in finding solutions to real-world problems. Through the proposed knowledge analysis methodology, scientists can learn about livestock keepers’ reasoning and problem-solving abilities, but also about aspects of the production processes where control is weak and new control possibilities are sought after. The analysis also yields information on how livestock keepers try to achieve their production goals despite restrictions and disturbances given by the production environment. This understanding is of high importance in low-external-input systems that have little scope to control production conditions but that do adapt to them.