The considerable progress in qualitative methods is directly connected with developments in the field of text-interpretation. On the basis of a thorough reconstruction of their formal structures texts are treated as autonomous domains of self-referential systems. Such a methodological status has been denied to pictures in empirical research in the field of social sciences up until now. The documentary method, based on Karl Mannheim's Sociology of Knowledge, opens up methodical access to pictures. Methodologies from art history (Panofsky, Imdahl) can thus become relevant for empirical research in social sciences. Connections to semiotics (Barthes, Eco) and philosophy (Foucault) are worked out in their consequences for qualitative methods. Thus verbal contextual and preknowledge can be controlled methodically in the documentary interpretation of pictures. The reconstruction of formal structure of pictures becomes of central importance in analysis. All of this will be demonstrated by examples from research practice.