In public administration, attempts to specify a unified interpretation of law seem to end in a specification with little operational meaning at all. The same unit of discourse in the sources of law usually plays many different knowledge roles, and ends up with – sometimes subtly – different operational meanings in each.

Knowledge acquisition from law in public administration usually subscribes to the notion of tasks to express the use of knowledge, even though it has clear deficiencies in dealing with variations in meaning due to context, particularly when dealing with a social construct like the legal institution. In the academic field we on the other hand see a move to multi-agent systems to explain the meaning of legal institutions as described by the sources of law. These however hardly do justice to the wide variety of problem solving behaviours found in the organization, and the pragmatic reasons for that variety.

In this paper we make an inventory of generic problem solving tasks in public administration, based on our experiences in case studies in a tax administration and an immigration and naturalization administration. We propose a typology of problems and discuss its conceptual connection to social agent roles that can be simulated in a multi-agent simulation environment.