Geography Reference
In-Depth Information
time is an unambiguous historical dimension. The case is similar in economic geogra-
phy, which can take up from its old position in its 'historical' research orientation. The
historical viewpoint, however, often suf ers from the weakness that it examines complex
interrelations less in an analytical manner, and instead adheres to a descriptive and phe-
nomenological viewpoint. Therefore, a co-evolutionary approach would require a more
detailed analysis of the varying temporality of the co-evolving populations. With the
concept of 'selective historical dependence', Setteri eld (1995, p. 22f.) calls for a selection
of the temporal context but at the same time warns of the dangers of 'historical deter-
minism'. The extent of the temporal dimension is determined by the object of research
and its scale: long-term processes, for example, are involved in the co-evolution of i rm
populations, short-term processes in the development of routines and practices in i rms.
Co-evolution in the narrower sense can only exist if processes take place more or less
simultaneously and interdependently; it is neither mere 'contingent' nor 'context' nor an
accompanying or subsequent event. This does not mean, however, that co-evolutionary
processes must therefore always take place at the same pace and over the same lifespan.
Martin and Sunley (2006) have rightly pointed out the positive lock-ins that are abso-
lutely essential in co-evolution, which make a sectoral path or a regional path possible.
Institutional co-evolution brings about the breaking of traditional institutions such as
traditional sectors, rules, or regions in favour of enabling new technologies, sectors, rules
and regions. Later lock-ins run the risk of disadvantageous contraction. This shifting of
characteristics poses questions with regard to historically specii c events: How does co-
evolution break down and what does this mean for the sectors and regions af ected (see
Schamp, 2005; also see Hassink, Chapter 21 in this volume)?
Co-evolution in the sense proposed here therefore only exists under three conditions:
(1) it must be possible to dif erentiate between two populations that (2) more or less
simultaneously (3) have a reciprocal causality. However, other disciplines have dif erent,
sometimes 'softer' concepts of co-evolution. For example, Dopfer and Potts (2004) refer
in their evolutionary realism approach to the 'bi-modality' and synchronicity of a popu-
lation of rules and of a population of supporting agents in order to account for evolution
(not co-evolution!). In contrast, in organisation science, business models resulting from
a specii c searching behaviour among new industries are also seen as co-evolving (e.g.
Huygens et al., 2001). Where are the boundaries of the concept of co-evolution? This
chapter would argue here for a stringent interpretation of the concept. It is in this sense
that the following sections examine the concept of co-evolution as applicable to tradi-
tional subjects of empirical research in economic geography.
3. The co-evolution of two i rm populations
For some classical economic geographers, it was self-evident that certain industrial
sectors developed jointly in the eighteenth and nineteenth centuries, for example coal
and steel or textiles/shoes and mechanical engineering. Geographers have discussed this
issue for the twentieth century with a less dei nite outcome, for example plastic injection
moulding and toolmaking or mail order business and related services (see Schulz et al.,
2006) or internet providers and venture capital i rms (see Zook, 2002). When economic
geographers or economic historians addressed the emergence of linked industrial sectors,
this usually involved little more than a historical narrative. There was no general analy-
sis of the mechanisms of evolutionary interaction. Economic geographers were only
