Geography Reference
In-Depth Information
development of appropriate computational architectures (such as multi-agent models
and dynamical systems models), and instead urge a more philosophically inclined social-
ontological approach. What precisely does it mean to talk of the economic landscape as
a complex system? In what sense is the economic landscape a meaningful complex system
to which the concepts of complexity thinking can be meaningfully applied? What does
connectivity mean and how do we distinguish partial from strong connections? These
are dii cult questions. To be analytically useful, complexity is not something that just
bolts on to or can be blended with an existing conceptual/theoretical framework to add
a 'complexity perspective' or 'evolutionary perspective'. Nor is it sui cient to invoke the
terminology and concepts of complexity science without thinking through what these
concepts are being applied to, and what they mean in an economic-geographical context.
Challenging questions though these are, in the view of Martin and Sunley the answers
could well be rewarding.
A way of dealing with complex systems from an evolutionary perspective is to analyse
how networks of agents evolve over time. While the study on network evolution is still
in a premature phase (see for example Powell et al., 2005), there is growing interest from
researchers to employ social network tools to describe and explain the evolution of
network structures and their performance over time. Economic geographers have started
to contribute to this emerging body of literature only quite recently. They are increas-
ingly aware that knowledge networks, and their spatial coni guration, play a crucial role
in the innovation process, and therefore may be considered a driving force of the evolu-
tion of the economic landscape.
However, network analysis is still very much underde-
veloped in the geography of innovation, and this is certainly true for an evolutionary
approach to this topic, though work is beginning to emerge (see, for example, Giuliani,
2007; Glückler, 2007; and some of contributions to this handbook, such as Breschi et al.,
Cantner and Graf, Giuliani, and Glückler; Chapters 16, 17, 12 and 14 respectively).
In Chapter 5, Boschma and Frenken take up this challenge by proposing an evolution-
ary perspective on the spatial evolution of innovation networks. They draw on insights
from the proximity literature (Boschma, 2005; Torre and Rallet, 2005) to explain the
evolution of the structure and performance of networks. Boschma and Frenken con-
ceive dif erent forms of proximity as alternative driving forces of network formation,
geographical proximity being one of them. They propose an evolutionary perspective
on the geography of network formation that is i rmly embedded in a proximity frame-
work (see also Sorenson et al., Chapter 15). Boschma and Frenken claim that proximity
may be considered a prerequisite for agents to connect with and to enhance knowledge
spillovers, but proximity between agents does not necessarily increase their innovative
performance, and may possibly even hinder it. The authors refer to this as the 'proxim-
ity paradox'. Boschma and Frenken then turn to the long-term dynamics of networks,
and discuss how such dynamics may be related to the changing role of proximity in the
formation and performance of innovation networks. In this respect, crucial questions
are how far and in what ways dif erent proximities induce path dependence in the spatial
evolution of networks (see, for example, Glückler, 2007; Ter Wal, 2009), and how this
process depends on spatial context. This is an area of research that is still strongly under-
developed, though considered crucial for the further development of an evolutionary
perspective on the spatial evolution of networks. According to Boschma and Frenken,
the ultimate goal is to develop a dynamic network approach that also accounts for the
