Intelligent World Model

In our approach we see a social system as a concept-processing, purpose-oriented „actor” interacting with its environment (which may contain other such actors), A brief survey follows; a fuller account is given elsewhere.

The concept-processing actor
Concepts are information patterns which carry a certain meaning for the actor, eg „the house is on fire”, „environmental pollution”, „democracy”. His actions are understood as responses to the perception, recognition, and processing of concepts, with the aid of other concepts stored in his memory. This view goes beyond the classical stimulus-response type of unconditional, unreflected information processing, although the underlying sequence is similar. We believe that the behaviour of a social system can be approximated as the interaction of its dominant actors. In a simulation the choice of relevant actors is determined by the task itself, which also determines the degree of aggregation or disaggregation. Individuals, collectivities, organisations, and institutions can all be viewed as actors. Each actor can be considered as a concept-processing system interacting with its environment. Its essential functions are performed by its physical system containing the perceptors and effectors as well as the material part of the concept-processing system. The programs and concepts necessary for the cognitive processes are contained in a memory. While some of them are concepts about factual objects and relationships, others concern normative concepts and relationships (discussed later). The sequence of concept-processing (which may or may not lead to observable behaviour) is sketched in Figure 1 and in more detail in Figure 2. Information about the state of the system and its environment is received by the perceptors and transformed into concepts. These concepts may or may not correspond to the real state. Pattern recognition and filtering play a role here, along with other factors: cognitive abilities, world knowledge, affects, and emotions. The concepts about the system state are compared to available normative „orientor” concepts about a desirable state, if any are available. If such guiding images (ideals, Leitbilder) are not available — eg in a situation not previously encountered or considered — then the likely contributions of the situation (as perceived) to the satisfactions of the system’s interests and orientation must be determined. This requires a mapping of the (perceived) situation on the system’s orientation space, and a corresponding evaluation. In both cases (the comparison of the perceived situation with existing ideals, or its mapping on the system’s orientation space), discrepancies may result which determine priorities and motivations in the subsequent problem solving and policy search. A corresponding behavioural plan is then developed and tested using the internal model. Concepts about possible and probable resulting states are projected by the internal model and are again analysed as to their possible effect on, or violation of, orientor concepts.