In Erkentniss als Konstruktion Luhmann formulates 5 conditions which one must pay attention to in production of (scientific) knowledge. The conditions are formulated in general terms. They rather pin points what must be keep in mind in empirical research, than prescribe strategies for construction and analysis of empirical information. A quick overview on empirical research drawing on Luhmann reveals a diverse complex of analytical strategies and empirical methods. Despite differences between strategies and methods they have in common that understanding of uttered information is crucial in their production of empirically founded knowledge. However research generally seems to pay more attention to production of uttered information than to selection of understanding. The aim of this contribution is to sketch out a suggestion to how selection of understanding can be systematized in order to produce enhanced transparency in selection of understanding as well as enhanced sensitivity and definition in dept. The contribution suggest that we distinguish between three types of understanding; simple, complex and hyper-complex understanding. Simple understanding is the simultaneous selection of understanding which accomplish each single unit of communication. From this perspective simple understanding can be described as a first-order observation. Complex understanding observes the difference used in selection of simple understanding. In this perspective complex understanding can be described as a second-order observation. Hyper-complex, or third-order observation, observes conditions for indication; that is conditions for actualization of the specific form of the difference constructed through second-order observation. The three types are partly deduced from Luhmanns concept of understanding, partly generated through analyses of empirical data.