Chris Dillon, The Open University, UK

Engineers use a variety of models to understand the systems they design and build. Developments in areas such as communications, control and switching systems, notably those moving into the public domain after the end of World War II, increasingly used models drawing on complex mathematical ideas. However, rather than working directly with the mathematics, engineers developed ways of describing systems by constructing models that represent behaviour pictorially. This paper explores the idea that the models engineers developed and used were not just convenient representations of systems or their behaviour, rather the pictorial representations provided new ways of talking that became part of the language of a community of practice.

The idea of a community of practice, introduced by Etienne Wenger and Jean Lave in the context of social theories of learning and subsequently further developed by Wenger, has gained currency in the study of how groups of practitioners learn and create meaning for themselves. For Wenger the key features of a community of practice are captured by a domain of expertise, a community of practitioners that share and learn from each other, and a practice that has ‘experiences, stories and tools’ in common.  From this perspective the development of engineering design techniques as they emerged in the late 1940s and early 1950s offers insights into the roles that models played in the mediation of ideas within the community of control and communications engineers.

Mary Morgan and Margaret Morrison, editors of the collection of papers ‘Models as Mediators’, suggest that it is because scientific models are ‘…neither one thing nor the other, neither theory nor data, but typically involve some of both […], that they can mediate between theory and the world’. In this paper I want to suggest that, for engineering communities focused on design as well as analysis, this view of models as mediators should be enhanced to explicitly include the community of practice in which the model is developed and embedded. More specifically, models of engineered systems’ behaviour should be understood not only as linking theory and the real world of designed objects, but also as playing an important role in mediating understanding, learning and conversation within the community.

To explore this I shall use examples drawn from the emergence of two distinct types of model: frequency-domain techniques in control engineering, and mapping techniques in the design of switching systems. In the former, graphical techniques based on the 1930s work of Nyquist, Black and Bode at the Bell Telephone Laboratories were used extensively to replace differential equations and algebra in control systems design problems from precision servomechanisms to large scale process plant. In the latter, the mathematics of Boolean algebra was simplified by a diagrammatic representation, due to Veitch (1952) and adapted by Karnaugh (1953), to represent the logical switching functions required in telephone networks, sequence control of industrial processes and the embryonic digital computer industry.

New models, tools and techniques take time to become assimilated and embedded sufficiently deeply in a community to ensure common conventions and ways of talking. The evolving discourses in the communication and control engineering communities following World War II offer historians of technology a view of this process in action. The coming together of ideas and practices from Europe and the United States created an environment in which informal learning between practitioners fuelled new ways of talking about systems. I want to suggest that it is the role of mediating models in this process that contributed to new shared understandings in the engineering community of practice.