Conceptual Modeling in Social and Physical Contexts

The history of the computing sciences shows a shift in attention from the syntactic properties of computation to the semantics of computing in the real world. A large part of this shift has been brought about by the introduction of conceptual modeling languages. In this paper I review this history from the early 1970s and identify the elements of real-world semantics that these notations have been used for. In the physical domains typical of control systems, conceptual modeling is always combined with causal modeling in order to register and control behavior in the domain. Because causal relationships are domain-specific, conceptual modeling languages in physical domains can be expected to evolve into domain-specific languages used by engineers. By contrast, in social domains causal modeling plays a minor role. In social domains conceptual models are shared by the people in the domain, and therefore constitute the domain. This creates a different mechanism for registration and control, in which events can be made to occur by means of social convention. Because conceptual models constitute the social world, we can expect conceptual modeling languages to evolve into domain specific languages here too, but in contrast to conceptual modeling languages in physical domains, they will be used as means of communication between engineers and members of the social domain. This paper ends with a plea for more specialization and less standardization in conceptual modeling.