Based on an extensive analysis of a course in process automation engineering we explore the current state of research on complex learning by means of a request analysis. The analysis compared the type of research on complex learning which is needed with the actual state of the art of learning and the current relevant approaches we find today. Research on complex learning has to answer at least three questions: (1) What is a useful definition of learning? (2) What changes - how to describe what is learnt? (3) How does it change - how to describe the process of learning? We try to answer these three questions by seven theses, which describe important features of future research on complex learning.