Innovation - the endless frontier: historical modes of engineering science

In this article we argue that knowledge production in industrial settings does not follow neat schemata, such as the linear model or the Mode 1/Mode 2 dichotomy. Rather than moving from research to market via development, innovation and production, it is an iterative process where information and knowledge move back and forth between different actors and milieus. Engineering science is a many-sided practice carried out at different institutions and is largely driven by a need for continuous novelty. By means of a detailed historical case study of industrial R&D in the area of engine design we document how theoretical knowledge encountered on-the-spot observations, how well educated engineers met ordinary bus drivers and how engineering test procedures clashed with economic interests. Our story line is about 'crossover' processes, trial-and-error 'research', recursive development loops, and flexibly acting innovation networks.