Dr. François Lafond

The evolution of knowledge systems

Year: 2014

Robin Cowan

The creation and allocation of economic wealth ultimately relies on the creation and diffusion of knowledge. As a result, understanding the dynamics, organization and viability of economies requires in-depth analysis of knowledge systems. This thesis proposes to study knowledge systems as self-organizing two-mode networks. Two-mode networks have two types of nodes, and the links are only between nodes of different types. These are self-organizing in the sense that simple rules of evolution lead to a rich but patterned dynamics. The thesis builds on the literature on social (agent-agent) and epistemic (idea-idea) networks to study socio-epistemic co-evolution (agent-idea networks). It is found that: (i) stable power law distributions of ideas' popularity naturally emerge from innovation and face-to-face diffusion; (ii) this dynamic is compatible with the observed (shifted) power law distribution of citations, and (iii) the generalized beta size-rank relation observed for patent classes can be explained by a slowdown in the growth of the number of classes. A general lesson from this work is that knowledge systems often exhibit non-equilibrium and non-linear dynamics, which may cast doubts on their long term viability.