By innovation, we do not mean only the invention of new kinds of artifacts. We refer to the processes through which new artifacts are conceived, designed, produced and integrated into patterns of use. These processes necessarily involve the construction of new patterns of interaction among agents, and hence transformations in the organization of what we call agent space.
Thus there is an inextricable linkage between the dynamics of change in the space of artifacts and in the space of agents. These dynamics are mediated by the way in which the relevant agents represent the contexts in which they act: in particular, their attributions about the identity of the other agents with whom they interact and the functionality of the artifacts around which their interactions are organized.
But in such a fast changing environment everything – agents, artifacts, attributions – is permanently under construction. So participants in innovation processes confront ontological uncertainty: not only can they not know for sure what the effects of their actions will be, but even the identities of the other agents whose actions will mediate those effects, and the criteria by which they will evaluate what such effects are worth when they happen is unknown when they must act.
In the Innovation Society, innovations occur in cascades: new artifacts, organizational transformations and new attributions of functionality are linked in a positive feedback dynamic.
To describe this dynamic, we distinguish two kinds of invention activities: those that are intended to deliver an existing functionality “better-faster-cheaper” than the artifacts that currently do so, and those that are designed to deliver new kinds of functionality. An innovation cascade can be initiated by either type of invention, and in any cascade both types are present. The positive feedback, though, runs primarily through the second kind, as follows:
The cascades that result from these positive feedback dynamics, characterized as they are by the generation of new attributions and the emergence of new patterns of agent interaction, are anything but linear and predictable.
Thus, the innovation literatures that concentrate on such phenomena as technological trajectories, supposed artifact-space dynamic regularities like Moore’s Law, passive adoption processes in the course of which neither artifacts, organizations or attributions change (innovation diffusion), or “trickle-down” pathways that lead from scientific breakthroughs in science to surprise-free developments in applied science to engineering advances to new products – all these have little of interest to say about the innovation cascades in agent-artifact space that we believe lie at the heart of innovation society dynamics.